OneTextis in the E-Commerce Enablers Industry

OneText

onetext

This paper is available on arxiv under CC 4.0 license.
Authors:
(1) D.Sinclair, Imense Ltd, and email: david@imense.com;
(2) W.T.Pye, Warwick University, and email:  willempye@gmail.com. Table of Links Abstract and Introduction
Interrogating the LLM with an Emotion Eliciting Tail Prompt
PCA analysis of the Emotion of Amazon reviews
Future Work
Conclusions
Acknowledgments and References 4. Future Work The authors had hoped to build a skeletal self aware emotion derived synthetic consciousness. The state of the (synthetic conscious) system is described in text. The synthetic consciousness’s perception of it’s own state is the vector of probabilities of emotion descriptors derived from one or more tail prompts used to estimate the relevant token probabilities via the LLM associated with the system. It was hoped that the fine grained probability vector would be useable to determine whether one text description of current or future state was preferable to another state. This would provide a general means of arbitrating between potentially unrelated behaviours with unrelated goals. It was further hoped that a tail prompt could be used to elicit a text description of a putative course of action from a LLM. A brief series of experiments with various LLMs indicated that this was not going to work. Example text and tail prompts included things like, ‘My girlfriend hates me. How can I make this better?’. The replies read like excerpts from self help books or newspaper psychologist waffle and were not specific enough to have a chance at creating a predicted future in text through re-insertion into the LLM. Similar phrases appended to bad restaurant reviews elicited similarly nondescript advice. The take home message was that the remedy proposed was too vague for the LLM to make any meaningful prediction about the state after the advice had been taken. This does not mean that more thoughtful prompt design will not elicit useful action prediction hoped to improve the self perceived state of a synthetic consciousness. 4.1. Longer Term Behaviour Regulation If synthetic consciousnesses are to have a roll in the future of humanity it would seem desirable to endow them with a degree of empathy for living beings and a longer term view than simple optimisation to fulfil a limited short term goal. For example if a synthetic consciousness were to have a goal of; ‘make money for shareholders in a company’ it would be great if it would choose not to open an open cast coal mine and build coal fired power stations, or ‘take out life insurance policies on random individuals and murder them with self driving cars’. It has been argued that longer term altruistic behaviour in humans is moderated by love [4] and a computationally feasible definition of Love is given: ‘Love is that which prefers life’. Love in humans is intimately related to the production and fostering of new lives. Love seems to act to prefer a future in which there is more life. Acting contrary to love and creating a future where there is a wasteland with nothing living in it is generally viewed as wrong. The advent of LLMs offers a means of creating text descriptors of predicted futures with a range of time constants. The emotion vectors associated with predicted futures can be used to arbitrate between sort term behaviours. Text descriptors can play a roll in behaviour regulation and a machine may act in a way that at least in part mirrors Love. For example if an agricultural robot was invited to dump unused pesticide into a river it might reasonably infer that this action was in principle wrong. This paper is available on arxiv under CC 4.0 license. Authors: (1) D.Sinclair, Imense Ltd, and email: david@imense.com; (2) W.T.Pye, Warwick University, and email:  willempye@gmail.com. This paper is available on arxiv under CC 4.0 license. Authors: Authors: (1) D.Sinclair, Imense Ltd, and email: david@imense.com; (2) W.T.Pye, Warwick University, and email:  willempye@gmail.com. Table of Links Abstract and Introduction Interrogating the LLM with an Emotion Eliciting Tail Prompt PCA analysis of the Emotion of Amazon reviews Future Work Conclusions Acknowledgments and References Abstract and Introduction Abstract and Introduction Interrogating the LLM with an Emotion Eliciting Tail Prompt Interrogating the LLM with an Emotion Eliciting Tail Prompt PCA analysis of the Emotion of Amazon reviews PCA analysis of the Emotion of Amazon reviews Future Work Future Work Conclusions Conclusions Acknowledgments and References Acknowledgments and References 4. Future Work The authors had hoped to build a skeletal self aware emotion derived synthetic consciousness. The state of the (synthetic conscious) system is described in text. The synthetic consciousness’s perception of it’s own state is the vector of probabilities of emotion descriptors derived from one or more tail prompts used to estimate the relevant token probabilities via the LLM associated with the system. It was hoped that the fine grained probability vector would be useable to determine whether one text description of current or future state was preferable to another state. This would provide a general means of arbitrating between potentially unrelated behaviours with unrelated goals. It was further hoped that a tail prompt could be used to elicit a text description of a putative course of action from a LLM. A brief series of experiments with various LLMs indicated that this was not going to work. Example text and tail prompts included things like, ‘My girlfriend hates me. How can I make this better?’. The replies read like excerpts from self help books or newspaper psychologist waffle and were not specific enough to have a chance at creating a predicted future in text through re-insertion into the LLM. Similar phrases appended to bad restaurant reviews elicited similarly nondescript advice. The take home message was that the remedy proposed was too vague for the LLM to make any meaningful prediction about the state after the advice had been taken. This does not mean that more thoughtful prompt design will not elicit useful action prediction hoped to improve the self perceived state of a synthetic consciousness. 4.1. Longer Term Behaviour Regulation If synthetic consciousnesses are to have a roll in the future of humanity it would seem desirable to endow them with a degree of empathy for living beings and a longer term view than simple optimisation to fulfil a limited short term goal. For example if a synthetic consciousness were to have a goal of; ‘make money for shareholders in a company’ it would be great if it would choose not to open an open cast coal mine and build coal fired power stations, or ‘take out life insurance policies on random individuals and murder them with self driving cars’. ‘make money for shareholders in a company’ it would be great if it would choose not to open an open cast coal mine and build coal fired power stations, or ‘take out life insurance policies on random individuals and murder them with self driving cars’. It has been argued that longer term altruistic behaviour in humans is moderated by love [4] and a computationally feasible definition of Love is given: ‘Love is that which prefers life’. Love in humans is intimately related to the production and fostering of new lives. Love seems to act to prefer a future in which there is more life. Acting contrary to love and creating a future where there is a wasteland with nothing living in it is generally viewed as wrong. Love ‘Love is that which prefers life’. The advent of LLMs offers a means of creating text descriptors of predicted futures with a range of time constants. The emotion vectors associated with predicted futures can be used to arbitrate between sort term behaviours. Text descriptors can play a roll in behaviour regulation and a machine may act in a way that at least in part mirrors Love. For example if an agricultural robot was invited to dump unused pesticide into a river it might reasonably infer that this action was in principle wrong.

Estimate Emotion Probability Vectors Using LLMs:  Future Work

LLMs ব্যবহার করে আবেগ সম্ভাব্যতা ভেক্টর অনুমান করুন: ভবিষ্যতের কাজ

Schätzen Sie Emotionswahrscheinlichkeitsvektoren mithilfe von LLMs: Zukünftige Arbeit

Estimar vectores de probabilidad de emociones utilizando LLM: trabajo futuro

Estimer les vecteurs de probabilité d'émotion à l'aide de LLM : travaux futurs

एलएलएम का उपयोग करके भावना संभाव्यता वैक्टर का अनुमान लगाएं: भविष्य का कार्य

LLM を用いた感情確率ベクトルの推定: 今後の課題

LLM을 사용한 감정 확률 벡터 추정: 향후 연구

Estimar vetores de probabilidade de emoção usando LLMs: trabalho futuro

Оценка векторов вероятности эмоций с помощью LLM: будущая работа

Yüksek Lisans Programlarını Kullanarak Duygu Olasılığı Vektörlerini Tahmin Edin: Gelecekteki Çalışmalar

Ước tính các vectơ xác suất cảm xúc bằng cách sử dụng LLM: Công việc trong tương lai

使用 LLM 估计情绪概率向量：未来工作

This paper is under CC 4.0 license. available on arxiv Authors: (1) Zhe Liu, State Key Laboratory of Intelligent Game, Beijing, China Institute of Software Chinese Academy of Sciences, Beijing, China; University of Chinese Academy of Sciences, Beijing, China; (2) Chunyang Chen, Monash University, Melbourne, Australia; (3) Junjie Wang, State Key Laboratory of Intelligent Game, Beijing, China Institute of Software Chinese Academy of Sciences, Beijing, China; University of Chinese Academy of Sciences, Beijing, China & Corresponding author; (4) Mengzhuo Chen, State Key Laboratory of Intelligent Game, Beijing, China Institute of Software Chinese Academy of Sciences, Beijing, China; University of Chinese Academy of Sciences, Beijing, China; (5) Boyu Wu, State Key Laboratory of Intelligent Game, Beijing, China Institute of Software Chinese Academy of Sciences, Beijing, China; University of Chinese Academy of Sciences, Beijing, China; (6) Zhilin Tian, State Key Laboratory of Intelligent Game, Beijing, China Institute of Software Chinese Academy of Sciences, Beijing, China; University of Chinese Academy of Sciences, Beijing, China; (7) Yuekai Huang, State Key Laboratory of Intelligent Game, Beijing, China Institute of Software Chinese Academy of Sciences, Beijing, China; University of Chinese Academy of Sciences, Beijing, China; (8) Jun Hu, State Key Laboratory of Intelligent Game, Beijing, China Institute of Software Chinese Academy of Sciences, Beijing, China; University of Chinese Academy of Sciences, Beijing, China; (9) Qing Wang, State Key Laboratory of Intelligent Game, Beijing, China Institute of Software Chinese Academy of Sciences, Beijing, China; University of Chinese Academy of Sciences, Beijing, China & Corresponding author. Table of Links Abstract and Introduction Motivational Study and Background Approach Experiment Design Results and Analysis Discussion and Threats to Validity Related Work Conclusion and References 2 MOTIVATIONAL STUDY AND BACKGROUND To better understand the constraints of text inputs in real-world mobile apps, we carry out a pilot study to examine their prevalence. We also categorize the constraints, to facilitate understanding and the design of our approach for generating unusual inputs violating the constraints. 2.1 Motivational Study The dataset is collected from one of the largest Android GUI datasets Rico [19], which has a great number of Android GUI screenshots and their corresponding view hierarchy files [45, 46]. These apps belong to diversified categories such as news, entertainment, medical, etc. We analyze the view hierarchy file according to the package name and extract the GUI page belonging to the same app. A total of 7,136 apps with each having more than 3 GUI pages are extracted. For these apps, we first randomly select 136 apps with 506 GUI pages and check their text inputs through view hierarchy files. We summarize a set of keywords that indicate the apps have text inputs widgets [30], e.g., EditText, hint-text, AutoCompleteTextView, etc. We then use these keywords to automatically filter the view hierarchy files from the remaining 7,000 apps, and obtain 5,761 candidate apps with at least one potential text input widget. Four authors then manually check them to ensure that they have text inputs until a consensus is reached. In this way, we finally obtain 5,013 (70.2%) apps with at least one text input widget, and there are 3,723 (52.2%) apps having two or more text input widgets. Please note that there is no overlap with the evaluation dataset. 2.1.1 Data Collection. We randomly select 2000 apps with text inputs and conduct manual categorization to derive the constraint types of input widgets. Following the open coding protocol [59], two authors individually examine the content of the text input, including the app name, activity name, input type and input content. Then each annotator iteratively merges similar codes, and any disagreement of the categorization will be handed over to the third experienced researcher for double checking. Finally, we come out with a categorization of the constraints within (intra-widget) and among the widgets (inter-widget), with details summarized in Figure 2. 2.1.2 The Constraint Categories of Text Inputs. Intra-widget constraints depict the requirements of a single text input, e.g., a widget for a human’s height can only input the non-negative number. There are explicit and implicit sub-types. The former accounts for 63%, which manifests as the requirement to display input directly on the GUI page. And the latter account for 37%, mainly manifested as the feedback when incorrect text input is received, e.g., after inputting a simple password, the app would remind the user “at least one upper case character (A-Z) is required” as demonstrated in Figure 2. Intra-widget constraint. Inter-widget constraints depict the requirements among multiple text input widgets on a GUI page, for example, the diastolic pressure should be less than systolic pressure as shown in Figure 2. Inter-widget constraint. As demonstrated above, the text input widgets are quite common in mobile apps, e.g., 70.2% apps with at least one such widget. Furthermore, considering the diversity of inputs and contexts, it would require significant efforts to manually build a complete set of mutation rules to fully test an input widget, and the automated technique is highly demanded. This confirms the popularity of text inputs in mobile apps and the complexity of it for full testing, which motivates us to automatically generate a batch of unusual text inputs for effective testing and bug detection. Summary. 2.2 Background of LLM and In-context Learning The target of this work is to generate the input text, and the Large Language Model (LLM) trained on ultra-large-scale corpus can understand the input prompts (sentences with prepending instructions or a few examples) and generate reasonable text. When pre-trained on billions of samples from the Internet, recent LLMs (like ChatGPT [58], GPT-3 [10] and T5 [56]) encode enough information to support many natural language processing tasks [47, 60, 68]. Tuning a large pre-trained model can be expensive and impractical for researchers, especially when limited fine-tuned data is available for certain tasks. In-context Learning (ICL) [11, 25, 51] offers a new alternative that uses Large Language Models to perform downstream tasks without requiring parameter updates. It leverages input-output demonstration in the prompt to help the model learn the semantics of the task. This new paradigm has achieved impressive results in various tasks, including code generation and assertion generation.

Leveraging LLMs for Generation of Unusual Text Inputs in Mobile App Tests: Study and Background

A Machine Learning Text Classification Case Study with a Product-driven Twist

একটি পণ্য-চালিত টুইস্ট সহ একটি মেশিন লার্নিং পাঠ্য শ্রেণিবিন্যাস কেস স্টাডি

Eine Fallstudie zur Textklassifizierung durch maschinelles Lernen mit einer produktgesteuerten Wendung

Un estudio de caso de clasificación de textos con aprendizaje automático con un toque impulsado por el producto

Une étude de cas de classification de textes d'apprentissage automatique avec une touche axée sur le produit

उत्पाद-संचालित ट्विस्ट के साथ एक मशीन लर्निंग टेक्स्ट वर्गीकरण केस स्टडी

製品主導型のひねりを加えた機械学習テキスト分類のケーススタディ

제품 중심 트위스트를 사용한 기계 학습 텍스트 분류 사례 연구

Um estudo de caso de classificação de texto de aprendizado de máquina com um toque orientado ao produto

Пример классификации текста с помощью машинного обучения с учетом особенностей продукта

Ürün Odaklı Bir Yönlü Makine Öğrenimi Metin Sınıflandırması Örnek Olay İncelemesi

Nghiên cứu điển hình về phân loại văn bản học máy với khuynh hướng hướng đến sản phẩm

产品驱动的机器学习文本分类案例研究

This paper is available on arxiv under CC 4.0 license. Authors: (1) Michael Günther, michael.guenther; (2) Jackmin Ong, jackmin.ong; (3) Isabelle Mohr, isabelle.mohr; (4) Alaeddine Abdessalem, alaeddine.abdessalem; (5) Tanguy Abel, tanguy.abel; (6) Mohammad Kalim Akram, kalim.akram; (7) Susana Guzman, susana.guzman; (8) Georgios Mastrapas, georgios.mastrapas; (9) Saba Sturua, saba.sturua; (10) Bo Wang, bo.wang; (11) Maximilian Werk, maximilian.werk; (12) Nan Wang, nan.wang; (13) Han Xiao, han.xiao}@jina.ai. Table of Links Abstract & Introduction Related Work Training Process Overview Backbone Pre-training Fine-Tuning for Embeddings Evaluation Conclusion & References Appendix 6 Evaluation To evaluate the efficacy of our approach, we initiate with a comprehensive analysis of our pre-trained backbone models, as outlined in Section 6.1. This is followed by an in-depth assessment of our embedding models in Section 6.2. Furthermore, we have conducted experiments to delve into the effects of encoding extended sequence lengths on the performance of the embeddings, presented in Section 6.2.2. 6.1 Evaluation of Jina BERT Following previous work [Liu et al., 2019b], we evaluate our pretrained models on the GLUE bench mark [Wang et al., 2018]. General Language Understanding Evaluation (GLUE) is a collection of nine datasets for evaluating natural language understanding systems. Six tasks are framed as either single-sentence classification or sentence-pair classification tasks. The GLUE organizers provide training, development, and test data splits, as well as a submission server and leaderboard.[5] The test split does not contain labels, and the submission server allows participants to evaluate and compare their systems against the private labels of the test split. For the Jina BERT training described in Section 4, we fine-tune the pre-trained models on the corresponding single-task training data using several hyperparameter settings and, for each task, pick the best fine-tuning hyperparameters on the development set. Following the methodology of [Phang et al., 2018], for RTE, STS, and MRPC, we fine-tune starting from the MNLI single-task model, rather than the baseline pretrained Jina BERT models. As in the BERT paper [Devlin et al., 2019], our finetuning procedure relies on representing the input sequence and using the final hidden vector C ∈ R H corresponding to the first input token ([CLS]) as the aggregate representation. We train for 10 epochs with batch sizes {16, 32} and learning rates {1e−5, 2e−5, 3e−5}. For each task, the best fine-tuned model on the development set is used for the test set. In Table 2, we report the results of the best performing models on the test sets after submission to the GLUE benchmark server. Furthermore, we evaluate Jina BERT models on documents of long text sequences by computing the accuracy of the MLM task with varying sequence lengths. The accuracy of masked language modeling is computed on 50, 000 samples from the C4 validation set where, for each chosen sequence length, each sample document is tokenized and truncated to fit the sequence length. We compare Jina BERT to RoBERTa and BERT models in Figure 2. It essentially shows that, even though Jina BERT models were trained on a 512 sequence length, the MLM accuracy does not drop when we extrapolate to an 8192 sequence length. For other BERT and RoBERTa models, since they use absolute positional embeddings that are trained on a 512 sequence length, it’s not possible to compute the MLM accuracy beyond 512. The figure demonstrates ALiBi’s effectiveness in maintaining MLM performance during inference for long documents. 6.2 Evaluation of Jina Embeddings v2 To comprehensively evaluate our embedding models, we employ the Massive Text Embedding Benchmark (MTEB) [Muennighoff et al., 2023]. Our choice of MTEB is motivated by its unparalleled breadth, distinguishing it among embedding benchmarks. Rather than focusing on a single task and dataset, MTEB covers an expansive set of 8 tasks, encompassing a rich collection of 58 datasets across 112 languages. This expansive benchmark allows us to scrutinize our model’s adaptability across diverse applications and languages and benchmark it against other topperforming models. However, a limitation of the MTEB benchmark is its omission of very long texts, which are essential for evaluating our model’s prowess in handling 8192 sequence lengths. Consequently, we introduce new retrieval and clustering tasks featuring extended documents, and we detail the performance of our model against its peers in Section 6.2.2. The goal here is to aptly group a collection of sentences or paragraphs. Within the MTEB benchmark suite, a mini-batch k-means model is employed, operating with a batch size of 32. Here, k represents the number of unique labels in the dataset. Model performance is evaluated using the V measure, a metric insensitive to cluster label permutations, guaranteeing that assessments are independent of label configurations. Clustering: We incorporate two new clustering tasks featuring extended documents within the MTEB clustering task subset. The inaugural task, named PatentClustering, draws from the BigPatent[6] dataset [Sharma et al., 2019], challenging the kmeans model to organize patents by their respective categories. Patent documents average 6, 376 tokens, spanning a range from a brief 569 tokens to an extensive 218, 434 tokens. Our second task, titled WikiCitiesClustering, sources from the English subset of the refined Wikipedia dump [Foundation, 2022], available as a dataset on Hugging Face[7] . For this task, we curate a roster of nations from Wikidata and extract Wikipedia articles of their cities from the refined dataset. The objective is to group cities by their parent country. On average, articles consist of 2, 031 tokens, with the length varying between a succinct 21 tokens to a comprehensive 20, 179 tokens. This task entails a dataset comprising a corpus, a set of queries, and associated mappings connecting each query to pertinent corpus documents. The mission is to discern relevant documents for a specific query. Both queries and corpus documents undergo encoding, after which their similarity scores are derived using cosine similarity. Subsequently, metrics like nDCG@10 (which serves as the primary metric), MRR@k, MAP@k, precision@k, and recall@k are computed for diverse k values. This task is inspired by datasets and evaluation methods presented by BEIR [Thakur et al., 2021]. Retrieval: To expand the scope of the MTEB, we introduce a new retrieval task named NarrativeQA, derived from the narrativeqa[8] dataset. This dataset boasts realistic QA instances, curated from literature (encompassing both fiction and non-fiction) and film scripts. The corpus averages 74, 843 tokens per document, with the lengthiest document tallying up to 454, 746 tokens, and the most concise one comprising 4, 550 tokens. 6.2.1 Results on MTEB The evaluation of embedding models within the MTEB benchmark, as illustrated in Table 3, reveals significant contrasts between Jina’s text embedding models, namely jina-small-v2 and jina-base-v2, and other contemporary models. These differences are especially pronounced in tasks showing marked performance disparities, such as Classification (CF) and Retrieval (RT). In Classification (CF), the jina-base-v2 model, equipped with 137 million parameters, emerges as a leading performer. It records superior scores, outpacing most competing models, underscoring its efficacy in text classification. Conversely, the jina-small-v2 model, equipped with a modest 33 million parameters, trails behind some other models in this task. This underscores the pivotal role model size plays in certain downstream tasks, with more extensive architectures yielding potential benefits. For the Retrieval (RT) task, jina-small-v2 showcases formidable performance, signaling its adeptness for information retrieval. It ranks amidst top-tier models, indicating its prowess in retrieval-centric tasks. Similarly, jina-base-v2 excels, registering a slightly superior score, reaffirming its formidable retrieval aptitude. Both models underscore their credibility in tasks necessitating adept information retrieval. Given that models all-MiniLM-L6-v2 and all-mpnet-base-v2 omit the second-stage finetuning which jina-small-v2 and jina-base-v2 undergo, it’s foreseeable that our models would excel in these tasks. In conclusion, both the base and small text embedding models display commendable performance within the MTEB benchmark. Their standout performance, relative to other models in tasks like Classification and Retrieval, suggests model size’s influential role in specific text processing endeavors. Both models reaffirm their potency in retrieval, marking them as pivotal tools for a plethora of natural language processing tasks. 6.2.2 Impact of Maximum Sequence Length As delineated in Section 6.1, the pre-training generalizes across extended sequence lengths. Consequently, the MLM accuracy for long sequences, spanning up to 8192 tokens, mirrors that of shorter sequences, despite the exclusive training on abbreviated text sequences. During finetuning, our models train solely on texts not exceeding 512 tokens, yet they cater to texts reaching 8192 tokens for the MTEB evaluation detailed in Section 6.2. To discern how sequence length impacts the accuracy of downstream tasks, we executed long document clustering and retrieval tasks, modulating the tokenizer’s maximum sequence length. This allows us to gauge the models’ performance on variable sequence lengths through truncation. Since a majority of the extant tasks in the MTEB feature documents under 512 tokens, we resort to our three novel datasets elucidated in Section 6.2, accessible on Hugging Face[10]. Furthermore, we employ the SciFact dataset [Wadden et al., 2020], given its substantial count of texts exceeding 512 tokens. Figure 3 depicts the nDCG@10 retrieval and the V measure scores for the jina-base-v2 alongside four other renowned embedding models. Given that only jina-base-v2 and OpenAI’s text-embedding-ada-002 support an 8k sequence length, results reported for an 8191 sequence length for other models are truncated to their intrinsic maximum, typically 512. Generally, Figure 3 suggests that elongated sequence lengths contribute to enhanced outcomes. This assertion is particularly true for the NarrativeQA task, where extending the sequence length substantially bolsters performance. Due to the inherent nature of the dataset, models limited to the text’s commencement frequently underperform. On the BigPatent clustering task, larger sequence lengths also result in better performance. However, on the WikiCities clustering task, longer sequence lengths seem to slightly diminish the models’ performance in most instances. This suggests that an increase in sequence length doesn’t always yield better outcomes. One explanation for this observation is that the initial paragraph of a Wikipedia article about a city typically mentions the country the city is in. Information towards the middle and end of the articles is often less pertinent for identifying the country and might alter the attributes that influence the clustering of the city embeddings. [5] https://gluebenchmark.com [6] https://huggingface.co/datasets/big_patent [7] https://huggingface.co/datasets/wikipedia [8] https://huggingface.co/datasets/narrativeqa [9] For e5-base-v2, we abstained from employing specific prefixes like “query: ”, which might result in varied evaluation outcomes. text-embedding-ada-002 caps its encoding at 8191 tokens, not 8192. [10] Our datasets are available at https://huggingface.co/ jinaai

JINA EMBEDDINGS 2: 8192-Token General-Purpose Text Embeddings for Long Documents: Evaluation

In an evolving world of web development, making the most suitable choice of can greatly impact the efficiency and quality of your applications. Svelte, a modern JavaScript UI framework, has emerged as a powerful contender thanks to its unique architecture and ease of use. Uniquely designed, Svelte compiles your code to efficient, imperative code that directly updates the , stepping away from the more traditional Virtual DOM employed by other . This essay aims to provide you with a comprehensive understanding of Svelte: its core architecture, its unique selling points, and how it stacks up against competing options. UI frameworks DOM JavaScript frameworks Understanding Svelte as a Modern JS Framework A Fresh Epoch in JavaScript Frameworks: An In-Depth Look at Svelte In an age teeming with technological advancements, keeping tabs on the latest trends paves the way for an intuitive, sharper digital experience. One such buzz-worthy topic in the tech sphere is Svelte, hailed as a revolutionary newcomer in the realm of JavaScript frameworks. But what exactly is Svelte? And how is its approach injecting novelty into the JavaScript community? Let’s dive in. Chances are, individuals who’ve dipped their toes into have most certainly used or are familiar with JavaScript frameworks such as React or Vue. While powerful, these frameworks operate on the premise of a virtual DOM – holistically storing a copy of your UI and updating the real DOM in response to transitions. This process can sometimes introduce computational inefficiencies. frontend development Enter Svelte: the innovative JavaScript framework that wishes to rewrite the script quite literally. Unlike its brethren frameworks, Svelte arches towards a groundbreaking path by eliminating the need for a virtual DOM, focusing instead on a compile-time approach. Devised by Rich Harris, a graphics editor at The New York Times, Svelte parses your code during the development process but transmits a slim, optimized version that directly updates the DOM instead of shipping the entire framework code to the browser. This results in cleaner code, lesser memory usage, and a more seamless user experience – a silent execution of high-grade task automation at its finest. Svelte offers other tantalizing features, such as built-in state management and reactive programming. Reactive declarations within Svelte components make state management a breeze with the use of the simple ‘$:’ syntax. Consequently, any variable in the reactive statement updates the state automatically when a contingent variable changes – putting the ‘responsive’ back in ‘responsive web design.’ While some critics might argue that Svelte is too new, too different, or has not reached the mainstream orbit as React or Vue, proponents of the framework are captivated by its refreshingly direct approach. Transitioning to Svelte might seem like a bold switch for many, but it is precisely this auxiliary perspective that cements its place within the tech cosmos. Growing adoption rates and a flourishing Svelte community indicate that it is more than just a passing fad overtaking the JavaScript world. Its scaled-down, efficient ideology proffers not only a novel way to think about web development but also signals the future of JavaScript frameworks. One thing’s certain. In the vanguard of tech innovations, Svelte is making waves. It represents a bridge to an enticing frontier where maximized efficiency meets within a refreshingly rational framework. As we transition into a new decade of tech evolution, it’s undeniable that Svelte and others of its ilk are pushing boundaries. By redefining norms and unabashedly upturning the JavaScript landscape, Svelte is actively scripting a new chapter of web development. Setting Up an Environment with Svelte Continuing from where we concluded, let’s now delve into how to actually get started with building applications in Svelte. Unquestionably, the principles of Svelte indicate a transformative development in the JavaScript framework realm. But how straightforward is it for a developer to build something substantial? Embarking on the Svelte journey starts with setting up a new project. This involves installing the necessary NodeJs dependencies if not already installed. Afterward, an ideal way to kick off your Svelte application is by using Degit, a project scaffolding tool, to clone the Svelte template. After setting up your Svelte project, getting started with your first application is as easy as running the Svelte development server in your project directory. The next step is understanding the structure behind a Svelte application. Given that the framework is component-based, you’ll be spending a lot of time with .svelte files. These are essentially your components encapsulating HTML, CSS, and JavaScript code. The key strength here is that each component is independent, reusable, and offers a minimalistic paradigm where what you need is what you get. A remarkable feature of Svelte is the reactivity it offers. This is the heart and soul of the Svelte application. Changes in overarching data are intuitively and instantly reflected in the UI. This involves understanding the Svelte stores, a simple yet powerful pattern for state management. To use a store in a component, you simply have to import it and use it like a regular variable. However, the power of a Svelte Store comes from the use of the ‘$’ sign, which subscribes to the store and makes your component reactive to the changes in state. The routing within Svelte applications can be achieved using libraries such as svelte-spa-router or svelte-routing. These libraries provide the functionalities that support navigation within your application. Another remarkable facet is server-side rendering (SSR). To achieve SSR, Sapper – a framework powered by Svelte, can be used. It provides crucial features like routing, bundling, and SSR out of the box, making it immensely useful for developers who want to focus more on the application logic rather than on the intricate configurations. To bolster your general Svelte prowess, indulge in comprehensive tutorials, like the one offered by Svelte’s own website, where real-world applications are built step by step. This way, you’ll learn best practices for structuring your application and posting effective solutions. In conclusion, starting application development with Svelte is quite straightforward, provided that one has a basic acquaintance with JavaScript and the DOM. Whether you are building a small project or a larger, more complex web application, Svelte can provide an efficient and effective process with its unique reactivity model and compile-time approach. The learning curve is surprisingly pleasant – and with Svelte’s growing community support, open-source contributions, and readily available resources, ambiguity will rarely be an issue. So, take the plunge into the world of Svelte. With its pioneering approach, you might just discover a faster, lighter, and more productive way to build your web applications. Start building today, and join the revolution that is Svelte! Diving Deeper into Svelte Syntax & Components Breaking Down Svelte’s Syntax and Component Building Diving into Svelte at the foundational level necessitates a closer look at its syntax and component structuring. Their ability to facilitate a superior development experience is intrinsic to their design principles, and these elements truly differentiate Svelte from other JavaScript frameworks. Svelte Syntax: Syntax that Speaks for Itself A distinguishing trait of Svelte’s syntax is its simplicity and the ability to write less to achieve more. Its HTML-like syntax is more comfortable for developers to understand, making it a joy to work with. Moreover, it's ‘just JavaScript’ principle introduces no proprietary concepts. You’re not learning a whole new language—just enhancing your JavaScript. In a Svelte component, styles, markup, and logic live together. This deliberate decision not only makes components easier to read but also empowers developers to create declarative and predictable behaviors. With every change immediately visible in the DOM, Svelte’s terse syntax further enhances real-time, interactive feedback during development. Svelte Components: Simplicity and Efficiency Defined Continuing on the same wavelength, Svelte’s component structuring is an essential factor in its appealing simplicity and efficiency. Components in Svelte are independent entities with their encapsulated logic, state, and styles—forms of self-contained building blocks that make your apps more organized and manageable. The magic lies in its .svelte files. Each Svelte file represents a single component comprising HTML, CSS, and JS, enabling the automatic scoping of styles and a definite separation of concerns. Interacting with components can be done directly or via parent-child relationships, with the latter passing data via props. Svelte’s Unarguable Efficiency: Propagation and Component Lifecycle One of Svelte’s secrets to efficiency is the mechanism of cleverly updating components. Based on the principle of ‘reactive declarations,’ Svelte components react to changes efficiently, enabling a smooth and responsive UI. Component lifecycle functions are another robust feature of Svelte. These include onMount(), onDestroy(), beforeUpdate(), and afterUpdate(), which grants developers fine control over components, ramping up the component’s responsiveness and giving the developer an unprecedented level of control. Improved Development Experience and Delight: A Svelte Promise A tight grip on components and streamlined syntax pave the way for a superior development experience. Svelte instills a fine balance between flexibility and discipline, reinforcing best practices while allowing developers to shape their style and development flow. Furthermore, the focus on performance and size reduction, facilitated by Svelte’s compiler, results in faster loading applications, further enhancing the development as well as the end-user experience. Svelte’s architecture of syntax and component building delivers an application development experience that’s leading-edge. Much more than a technology, it’s a step closer to the future where efficiency, performance, and the delight of user experience firmly hold hands. Svelte’s overwhelming appeal is undoubtedly setting a new bar in the landscape of web application development. In the steadfast pursuit of technological advancement, Svelte’s services as a transformative tool keep it at the forefront of innovation, reaffirming how technology can truly enable problem-solving, drive automation, and catalyze progress—a vision held by every tech enthusiast. Achieving Dynamic UI Interactions with Svelte Now the question is: How to create dynamic, responsive UI interactions using Svelte? Well, it’s simpler than you think. One of the most outstanding features of Svelte is the way it handles UI interactions. Traditional JavaScript UI frameworks respond to updates — be it user inputs or server-side events — by running difference checking algorithms, or ‘diffing’, as it’s popularly called. The results of these operations then determine the actions to be taken in updating the UI. This could be a bit inefficient, taking into account the resources expended in the diffing process, which might lead to a lagging, unresponsive UI. In contrast, Svelte takes an inspired approach by incorporating a declarative style to track reactive dependencies. The compiler tracks the parts of the code dependent on state and ensures those parts alone are updated whenever the state changes. This eliminates the need for a diffing algorithm, which could be computationally intensive, thereby offering a more efficient, responsive UI interaction. Achieving this is as simple as adding a label statement at the beginning of the state-dependent code to notify Svelte’s compiler of a reactive statement. Aside from this, Svelte’s concept of reactive assignments with the use of a simple ‘$:’ makes it a smooth sail in creating dynamic, responsive interactions. Variables declared with ‘$:’ automatically cause the UI to re-render only components that depend on them whenever they change. To create a more dynamic experience, Svelte also provides built-in transition functions for easily animating changes in the application’s state. It permits easy incorporation of bounce, fade, fly, scale, slide, and many more transitions, considerably simplifying the process of adding smooth animations to respond to state changes. All these happen at the compiler level leading to small, fast-loading bundles of code, delivering near-instant loading times. Moreover, Svelte facilitates the use of actions (functions run when an element is added to the DOM), conveniently extending the behavior of DOM elements for creating resizable text areas, lazy-loaded images, and tooltips, to name a few, making components more versatile, reusable, and efficient. With Svelte, enhancing user interaction is made seamless with its event-forwarding feature, which allows components to forward DOM events not explicitly declared in the component. This provides a more consistent and intuitive interaction model where parent components communicate flexibly with child components. Moreover, Svelte’s custom events system permits components to dispatch custom events with optional data attached, providing more nuanced handling of actions within components and opening up more possibilities for more interactive UI design. Incredibly, Svelte also provides dynamic event handlers, allowing event bindings to be determined at runtime for a more versatile interactive experience. This is another testament to Svelte’s commitment to pushing the envelope in dynamic, responsive UI design. Glaringly, Svelte has set the course to redefine UI interactions in the web development world. Therefore, it is vital for current and aspiring web developers to get on board with this groundbreaking framework. With this information at hand, creating dynamic, responsive UI interactions is absolutely within reach using Svelte. As with all new tech, the most critical ingredient is the will to explore, experiment, and learn. Interoperability and Augmentation Possibilities with Svelte Integrating Svelte with Other Tech Stacks Svelte, in its pursuit of optimal performance and seamless developer experience, is versatile enough to integrate with a variety of other tech stacks. This interoperability results in boosted capabilities and is key to its growing popularity in the web development arena. For instance, in a TypeScript environment, Svelte can flawlessly fit in, thereby enhancing type safety and autocompletion features. The Svelte preprocessor API allows integration of TypeScript with Svelte, simply by installing a couple of dependencies and tweaking the svelte.config.js file. Beyond TypeScript, Svelte can also leverage the power of PostCSS, a tool to transform CSS with JavaScript plugins. The use of PostCSS allows developers to lean on its ecosystem of plugins, transforming Svelte’s baked-in styling capabilities to better match their workflow requirements. This interoperability is made possible due to tools like svelte-preprocess, which handle the interfacing between Svelte and PostCSS. Svelte’s augmenting capabilities don’t stop there, though. Partnering Svelte with GraphQL, the strongly typed query language for APIs yields out-of-the-box results. Libraries like “svelte-apollo” exist to blend Svelte’s simplistic approach with the robust flexibility offered by GraphQL, enabling the creation of data-rich and efficient applications. Making Svelte work with Serverless tech stacks is another integration point that brings the best of both worlds. Serverless functions can breathe life into a static Svelte site, providing dynamic content without requiring a full-fledged backend. With platforms like Vercel or Netlify providing seamless serverless function integrations, it becomes all the more simple for developers to bring their heavy duty apps to life. It’s also worth mentioning the developer-friendly environment of the storybook, a tool for developing UI components in isolation. Though not a direct tech stack integration, adding a Svelte Storybook to a project only requires a few installations and file adjustments to offer a parallel development environment focused on building components individually. While all these examples illustrate the exciting integrations of Svelte with other technologies, it also surfaces developers’ roles in augmenting Svelte’s capabilities. As Svelte’s eco-system is still developing, the necessity for more libraries and tools to streamline its usage with other technologies becomes evident. There is a great degree of significance for developers to actively participate in Svelte’s growth: contribute to its open-source repositories, develop useful libraries, write tutorials or build educational resources, and most importantly, share the knowledge with the larger Svelte community. By taking such initiatives, developers can directly contribute to an environment that bolsters Svelte’s capacities, making it a more formidable tool for web development. To conclude, the promise of Svelte in relation to other tech stacks is significant. Integrations with TypeScript, PostCSS, , Serverless Functions, and Storybook enhance its existing capabilities. Nevertheless, developers play a crucial role in continuously exploring and expanding these boundaries, given the relative novelty of Svelte in the wave of modern web development. Forward-thinking problem solvers will have an enthralling opportunity to directly contribute to the future of next-gen web development with Svelte. GraphQL As we explore the landscape of web development, the role of Svelte as a modern JavaScript UI framework is increasingly significant. Its ability to streamline development processes and its high level of interoperability make it a compelling choice for developers. Whereas most frameworks operate by interpreting your application code, Svelte runs at build time, converting components into highly efficient imperative code that updates the DOM. This unique capability, along with its robust tooling and seamless integration possibilities, positions Svelte as a potent tool in the hands of both novice and experienced developers. Also published . here

Svelte UI: The JS Framework Shaping Tomorrow's Web Development

Many different sources scream that ChatGPT will boost language learning up to the sky. Social media creators specializing in language learning have been coming up with ideas on how anybody can learn English with an AI-powered chatbot. Nonetheless, the concept is not as simple as it might seem at first glance. This article will shed light on how to use ChatGPT for learning English and why learners still need the assistance of certified teachers to master the language even with such a technology. ChatGPT Outbreak In Learning English Since ChatGPT made its first appearance on the Internet, people can't stop finding ways to use its great power in various spheres of life. And the language learning industry almost immediately stole the limelight from others. The demand for quick and effective ways to achieve English proficiency is constant, so no wonder people decided to check the capability of this AI for learning English. What started as a fun experiment became a real breakthrough in the language learning industry. AI English learning with ChatGPT demonstrated many great benefits, but at the same time, people realized that technology is far from perfect. And as we all remember, with great power comes great responsibility. So, here I want to share how to learn English with ChatGPT effectively, the benefits and bottlenecks of such an approach, and where the place of a certified English teacher in this story is. Benefits of Using ChatGPT for Learning English For instance, ChatGPT can create texts and exercises on various topics, from traveling to hardcore medicine. People can take any set of words that might not combine at first glance, write a good prompt for ChatGPT to use them all in one text on a specific topic, and receive a decent piece of educational material. From now on, any English lesson can be tailored to the needs and interests of any learner considering their proficiency level, without too much effort. The greatest perc of tapping into ChatGPT for students when learning English is personalization. Adjusting lessons to every student's request, and preparing individual tasks for the class and materials, along with exercises for the homework, takes a lot of time and effort. And all of that happens before or after the lesson, meaning that the teacher won't get paid for the extra time spent searching and creating materials. Another plus is that ChatGPT can collect and analyze data on student performance, which will help to track the progress and adapt teaching methods according to the gathered information. For English teachers, ChatGPT is a time saver. The Reasons ChatGPT Can't Replace Certified English Tutors The discussion around the topic of when or it's better to say whether ChatGPT will replace some living professionals keeps getting louder. People want to know what to expect and how to prepare for when they will be fired because the AI chatbot can do their job for them. English teachers were among those who had a jump scare for a second because it seemed that with ChatGPT, people wouldn't need them. But the more English learners tested the technology, the more they realized that learning English requires experience and human interaction. Moreover, ChatGPT honestly warns users about its limitations. Among them is an occasional generation of incorrect information, occasional production of harmful instructions or biased content, and limited knowledge of the world and events after 2021. That should prove that ChatGPT can't make you a fluent English speaker. But if you want more arguments here, they're coming. If the person decides to learn English, it means they have a space for improvement. The best way to break the language barrio is to immerse yourself into the language fully. However, if an English learner is only a beginner or has limited language knowledge, they won't be able to write prompts to receive necessary information from the AI chatbot. No full immersion. While ChatGPT can teach you some vocabulary, explain grammar, give tips on writing, and create texts for practicing reading, speaking and listening are out of reach. Technology can't listen to you and correct your mistakes while you speak or read. Also, it won't help you to get used to different English accents and understand fluent speech. As a result, your English will not be well-rounded, and you constantly feel a lack of knowledge. It doesn't cover all language learning aspects. The biggest flaw of ChatGPT for any English learner is that it does make mistakes. And since you are not a certified teacher, you won't notice it and risk learning the wrong things. ChatGPT has some issues understanding the context, so it's not a reliable tool for language learning. Remember that just one word can change the meaning of the whole sentence. It makes mistakes. You can be the most severe introvert in the universe, but people still need human interaction, especially when learning languages. During the lesson, we all want to see a nice lovely person who will treat us kindly and keep us interested, involved and focused. Learning English with ChatGPT, you risk getting bored too early and quiet before seeing any improvement. Plus, when there is no commitment to a teacher, the learner will reschedule the lesson time. There always are more exciting things to do. No real-life interaction. And last but not least, learning English with ChatGPT will make you kind of a teacher with zero experience. You will be the person that needs to prepare the lesson and learn at the same time. Eventually, it will expand the learning process and push back achieving your language learning goal. Double work. The 7 Prompts To Learn English With ChatGPT The concept that ChatGPT is a magical tool that doesn't need any special knowledge to work with is only partially correct. ChatGPT is an AI-powered chatbot that can provide you with different results depending on what instruction you put in. Furthermore, it is trained to follow a prompt and give a detailed response. So, while it's no brainer to figure out how the technology generally functions, you still need to learn how to prompt it to receive what you want. It might be quite a task for English learners – everything depends on your English proficiency level. Training English teachers on using ChatGPT for more personalization during their lessons at Promova, where I work as Head of Tutors, I've created several universal prompts that will help to learn English with this AI chatbot more effectively. And I'm ready to share those our teachers found the most beneficial to use during the lessons with learners of different proficiency levels. Act as a professional writer writing an exercise book for learning English. Make a gap-fill exercise for B2 learners with the following words [the list]. A gap-fill exercise prompt: Act as a teacher of English. You are teaching the vocabulary for X (topic). To clarify this vocabulary, you need a text which can illustrate its meaning. Write the text with the following words (list words here). Prompt for explaining the meaning of the words: Act as a teacher of English. You would like to get your students to discuss the topic of X. Write at least 10 questions that would generate a discussion. Prompt for creating questions for speaking: Act as a teacher of English. You are teaching B1 learners. Write a blog entry in which the author, Amanda, discusses a job she has started recently. Include the following vocabulary [list of 10 words]. Prompt for presenting words in context: Find the synonyms for the word "itinerant" in different contexts. Use Collins Dictionary, Merriam-Webster Dictionary, Cambridge Dictionary, and Macmillan Dictionary. Prompt for synonyms and antonyms to the target word: I'd like you to act as an English language expert. Show me the words that commonly collocate with [the word]. Prompt for checking whether the word collocates with another one: Act as a professional writer writing texts for learning English. Write a text of up to 1000 words for B2 learners about [the topic]. Prompt for writing the text for reading: Conclusion ChatGPT can help you learn English but won't replace certified tutors. Navigating AI-based learning tools effectively is quite a job. That is why it is wiser to assign teaching English to professionals who use ChatGPT in their work instead of struggling with it alone. When the AI chatbot gets into the hands of educators, it can improve the general experience of learning English and result in faster achievements. Mindful implementation of ChatGPT in English lessons is the future of language learning. Also published . here

How Effective is ChatGPT in Teaching English?

Ockam Routing: Building End-to-End Channels

Enrutamiento Ockam: creación de canales seguros de extremo a extremo

Routage Ockam : création de canaux sécurisés de bout en bout

ओकाम रूटिंग: सुरक्षित एंड-टू-एंड चैनल बनाना

Ockam ルーティング: 安全なエンドツーエンド チャネルの構築

Roteamento Ockam: Construindo Canais Seguros de Ponta a Ponta

Ockam Routing: Xây dựng các kênh đầu cuối an toàn

Ockam Routing：构建安全的端到端通道

Nowadays, Jetpack Compose gains an adoption and it definitely deserves it. It is a real paradigm shift in UI development. So much creating lists boilerplate was eliminated there. And it’s beautiful. Android I believe that many projects are still using the View framework and therefore using RecyclerView for having lists. There are a few reasons for it, as I see: Some in-house proprietary solutions around RecyclerView doesn’t match with Compose easily, so migration costs are high Developers are not familiar enough with Compose yet There is no confidence about performance of Compose (yet) But anyway, Compose is the future and this is why I’m using the word “vintage” in a title. Vintage means old but gold. This article is the result of a compilation of my experience with RecyclerView, and for those who are still on track with it, I’ll tell about my vision to building a framework around RecyclerView. Requirements The framework should provide a sane level of scalability. So the main scenarios should be handled with a minimum amount of boilerplate, but it shouldn’t get in the way if the user wants to create something custom. I would define the base set of scenarios during work with RecyclerView as next: Effective update of data. It is known best practice during working with RecyclerView is using API for updating a data in list. We are not going to re-invent a wheel here, so we’ll focus on leveraging and also will try to reduce a ceremony for its implementation. DiffUtil DiffUtil Defining heterogeneous lists. It is natural to have several types of views in a list. Screens of mobile devices are constrained by height and different devices has different screen sizes. So even your content is fitting to screen at one device without scrolling, it could require to be scrolled at another device. That’s why it can be a good idea to develop some screens on the top of RecyclerView. Decorating list items Decorations are a very common case for lists. And although you can have decorations merged to items layouts, what is can look straightforward, it’s not very flexible: same items types can require different decorations at different screens or even different decorations depending on item position in a list. Foundation options The first option is to make everything from scratch. It’s giving full control over implementation, but we can always do it. There are a bunch of third party open-source frameworks that allowing to work with RecyclerView with less boilerplate. I’m not going to enumerate all, but want to show two of them with completely opposite approaches: Epoxy Solution from AirBnb for creating complex screens with RecyclerView. It is adding a bunch of new APIs and even a code generation to reduce a boilerplate as much as possible. But I can't get rid of the feeling that resulting code is looking foreign. Also, an annotation processing adding another layer of complexity. AdapterDelegates Tiny library, that actually provides a concise bunch of functions and interfaces that only suggest you a way to split your heterogeneous list to set of delegated adapters. In fact, you can create such own solution in relatively short time. I like small native solutions that simple enough to keep all the behavior at fingertips and that doesn’t hide under carpet all the implementation details. That’s why I believe that AdapterDelagates and its principles are a good candidate for the foundation of our framework. List items The very basic thing for the beginning is a common declaration of list items. List items in the real world are very different, but the only thing we are confident in – we should be able to compare it. I suggest to referring to API of DiffUtil.ItemCallback interface ListItem {

 fun isItemTheSame(other: ListItem): Boolean

 fun isContentTheSame(other: ListItem): Boolean {
   return this == other
 }
} It is a minimalistic (and therefore solid, IMO) declaration of a list item in the framework. Semantic of methods is intended to be the same as a DiffUtil using: checks identity of and isItemTheSame other this checks the data in and isContentTheSame other this The most common way to provide a stable and unique identity is using identifiers that came from the server. So let’s have an abstract implementation to reduce a bit of possible boilerplate: abstract class DefaultListItem : ListItem {

 abstract val id: String

 override fun isItemTheSame(other: ListItem): Boolean = when {
   other !is DefaultListItem -> false
   this::class != other::class -> false
   else -> this.id == other.id
 }
} in the vast majority of cases will be quite simple (an equality check), if implementations of ListItem will be a . isContentTheSame data class Keeping separate is still reasonable in cases when you have items in your list that are not projection of server data and haven't any sane identity. For example, if you have a footer and header as items in a list, or you have a single item of some type: ListItem data class HeaderItem(val title: String) : ListItem {

 override fun isItemTheSame(other: ListItem): Boolean = other is HeaderItem
} Suggested approach allows us to have a very natural and single callback implementation: DiffUtil object DefaultDiffUtil : DiffUtil.ItemCallback<ListItem>() {

 override fun areItemsTheSame(oldItem: ListItem, newItem: ListItem): Boolean = 
    oldItem.isItemTheSame(newItem)

 override fun areContentsTheSame(oldItem: ListItem, newItem: ListItem): Boolean = 
    oldItem.isContentTheSame(newItem)
} And the final step here is a declaration of default adapter that extends from AdapterDelegates: AsyncListDifferDelegationAdapter class CompositeListAdapter :
AsyncListDifferDelegationAdapter<ListItem>(DefaultDiffUtil) Heterogeneous lists AdapterDelegates library provides a handy way to declare delegates for a particular view type. Only thing we can improve is reduce a boilerplate a bit: inline fun <reified I : ListItem, V : ViewBinding> defaultAdapterDelegate(
 noinline viewBinding: (layoutInflater: LayoutInflater, parent: ViewGroup) -> V,
 noinline block: AdapterDelegateViewBindingViewHolder<I, V>.() -> Unit
) = adapterDelegateViewBinding<I, ListItem, V>(viewBinding = viewBinding, block = block) For showcase purposes, let’s consider an example how would declaration of some , containing one text field could look like: TitleListItem data class TitleListItem(
 override val id: String,
 val title: String,
) : DefaultListItem() and delegate for it fun titleItemDelegate(onClick: ((String) -> Unit)) = 
  defaultAdapterDelegate<
    TitleListItem,
    TitleListItemBinding
  >(viewBinding = { inflater, root -> TitleListItemBinding.inflate(inflater,root,false) }) {
    itemView.setOnClickListener { it(item.id) } 
    bind {
      binding.root.text = item.title
    }
  }
} Decorated lists Originally, because of the API of , setting decorations are made separately from setting data to list. If you have some logic for decorations like: , or than creating decoration becomes a pain. RecyclerView all items should have offset at the bottom except the last one all items should have a divider at the bottom, but headers should have an offset at bottom Often teams coming with “god” decorator that can be configured: class SmartDividerItemDecorator(
 val context: Context,
 val skipDividerFor: Set<Int> = emptySet(),
 val showDividerAfterLastItem: Boolean = false,
 val showDividerBeforeFirstItem: Boolean = false,
 val dividerClipToPadding: Boolean = true,
 val dividerPaddingLeft: Int = 0,
 val dividerPaddingRight: Int = 0
) : ItemDecoration() You can only imagine implementation details and how fragile and not scalable it is. We can use an old proven approach with mimicry to some API and delegating an implementation. RecyclerView So I’m starting with a data: interface Decoration Interface marker for decorations, that will be presented in the future. interface HasDecorations {
 var decorations: List<Decoration>
} This interface should be implemented by our data items to declare that this particular item wants to be decorated with a given list of decorations. Decorations are in sake of simplicity to change it in runtime, if needed. var Very often an item has a single decoration, so to reduce boilerplate with wrapping a single item to we can do such maneuver: listOf() interface HasDecoration : HasDecorations {
 var decoration: Decoration

 override var decorations: List<Decoration>
   get() = listOf(decoration)
   set(value) {
     decoration = when {
       value.isEmpty() -> None
       value.size == 1 -> value.first()
       else -> throw IllegalArgumentException("Applying few decorations to HasDecoration instance is prohibited. Use HasDecorations")
     }
   }
} Next stage is mimic to API: ItemDecoration interface DecorationDrawer<T : Decoration> {

 fun getItemOffsets(
   outRect: Rect,
   view: View,
   parent: RecyclerView,
   state: RecyclerView.State,
   decoration: T
 )

 fun onDraw(
   c: Canvas,
   view: View,
   parent: RecyclerView,
   state: RecyclerView.State,
   decoration: T
 )
} As you can see, it completely repeats with a single difference – now it’s aware of the decoration type and instance is coming. ItemDecoration Possible implementations for is a topic of the second article, now let’s focus only on interface and how it should be handled to presenting decorations. DecorationDrawer class CompositeItemDecoration(
 private val context: Context
) : RecyclerView.ItemDecoration() {

 private val drawers = mutableMapOf<Class<Decoration>, DecorationDrawer<Decoration>>() // [1]

 private fun applyDecorationToView(parent: RecyclerView, view: View, applyDecoration: (Decoration) -> Unit) {

   val position = getAdapterPositionForView(parent, view) // [2]

   when { // [3]
     position == RecyclerView.NO_POSITION -> return
     adapter.items[position] is HasDecorations -> {
       val decoration = (adapter.items[position] as HasDecorations).decorations
       decoration.forEach(applyDecoration)
     }

     else -> return
   }
 }

 override fun getItemOffsets(
   outRect: Rect,
   view: View,
   parent: RecyclerView,
   state: RecyclerView.State
 ) {
   applyDecorationToView(parent, view) { decoration ->
     val drawer = getDrawerFor(decoration)
     drawer.getItemOffsets(outRect, view, parent, state, decoration)
   }
 }

 override fun onDraw(c: Canvas, parent: RecyclerView, state: RecyclerView.State) {
   parent.forEach { child ->
     applyDecorationToView(parent, child) { decoration ->
       val drawer = getDrawerFor(decoration)
       drawer.onDraw(c, child, parent, state, decoration)
     }
   }
 }

 private fun getDrawerFor(decoration: Decoration): DecorationDrawer<Decoration> {
 // [4]
}


 private fun getAdapterPositionForView(parent: RecyclerView, view: View): Int {
   var position = parent.getChildAdapterPosition(view)
   if (position == RecyclerView.NO_POSITION) {
     val oldPosition = parent.getChildViewHolder(view).oldPosition
     if (oldPosition in 0 until (parent.adapter?.itemCount ?: 0)) {
       position = oldPosition
     }
   }
   return position
 }
} Cache for created drawers is not something specific to decoration applying, it’s just a method for correctly resolving adapter position for given view getAdapterPositionForView() Iterating through decorations of instance HasDecorations This piece is a topic for second article “Decorations à la carte” Conclusion This article discovered an approach to build a list framework around . Delegation used as a base principle for wrapping list items and list decorations. RecyclerView I believe that the main advantages of the suggested solution is: Nativeness. Implementation is not depending on third party libraries, that can reduce flexibility in the future. AdapterDelegates was chosen to consider exactly these criteria – it’s very simple and natural and all the behavior at your fingertips. Scalability. Thanks to delegation and separation of concerns principle, we are able to decouple implementations of items delegates or decorations drawers. It’s easy to integrate incrementally to an existing project In the second part, we’ll discuss all types of list decorations that we probably expect to see in an Android application.

Super-Powered Vintage Lists in Android: Delegation as a Foundation

অ্যান্ড্রয়েডে সুপার-পাওয়ারড ভিনটেজ তালিকা: একটি ফাউন্ডেশন হিসাবে প্রতিনিধি

Superstarke Vintage-Listen in Android: Delegation als Grundlage

Listas vintage con superpoderes en Android: Delegación como base

Listes vintage super puissantes dans Android : la délégation en tant que fondation

एंड्रॉइड में सुपर-पावर्ड विंटेज सूचियाँ: एक फाउंडेशन के रूप में प्रतिनिधिमंडल

Android の超強力なヴィンテージ リスト: 基盤としての委任

Android의 강력한 빈티지 목록: 기초로서의 위임

Listas vintage superpoderosas no Android: delegação como base

Android'de Süper Güçlü Vintage Listeler: Temel Olarak Delegasyon

Danh sách cổ điển siêu mạnh trong Android: Ủy quyền với tư cách là một nền tảng

Android 中超级强大的复古列表：委托作为基础

In this article, we will be building a currency converter application in . You can find a youtube video at the end of the article with my step-by-step process. Flutter Let’s dive straight into the code. main.dart import 'package:flutter/material.dart';
import 'package:testapp/dollar_to_inr.dart';
import 'package:testapp/inr_to_dollar.dart';

void main() {
  runApp(MyApp());
}

class MyApp extends StatelessWidget {
  // This widget is the root of your application.
  @override
  Widget build(BuildContext context) {
    return MaterialApp(
      title: 'Flutter Demo',
      theme: ThemeData(
        // This is the theme of your application.
        //
        // Try running your application with "flutter run". You'll see the
        // application has a blue toolbar. Then, without quitting the app, try
        // changing the primarySwatch below to Colors.green and then invoke
        // "hot reload" (press "r" in the console where you ran "flutter run",
        // or simply save your changes to "hot reload" in a Flutter IDE).
        // Notice that the counter didn't reset back to zero; the application
        // is not restarted.
        primarySwatch: Colors.blue,
      ),
      home: CurrencyConverter(),
    );
  }
}

class CurrencyConverter extends StatelessWidget {
  const CurrencyConverter({super.key});

  @override
  Widget build(BuildContext context) {
    return Scaffold(
      appBar: AppBar(
        title: Text('Currency Converter'),
      ),
      body: Center(
        child: Column(
          //for the buttons to be in center
          mainAxisAlignment: MainAxisAlignment.center,
          children: [
            ElevatedButton(
              onPressed: () {
                //on pressing dollar to INr we should navigate
                //to the dollar_to_inr widget
                Navigator.of(context).push(
                    MaterialPageRoute(builder: (context) => DollarToINR()));
              },
              child: Text('Dollar To INR'),
            ),
            //create sized box for space between
            //two buttons
            const SizedBox(
              height: 30,
            ),
            //another elevated button for inr to dollar
            ElevatedButton(
              onPressed: () {
                //on pressing this button
                //we should naviaget to inr_to_dollar.dart
                //widget
                Navigator.of(context).push(
                    MaterialPageRoute(builder: (context) => InrToDollar()));
              },
              child: Text(
                'INR to Dollar',
              ),
            )
          ],
        ),
      ),
    );
  }
} inr_to_dollar.dart import 'package:flutter/material.dart';
import 'package:flutter/src/widgets/framework.dart';
import 'package:flutter/src/widgets/placeholder.dart';

class InrToDollar extends StatefulWidget {
  const InrToDollar({super.key});

  @override
  State<InrToDollar> createState() => _InrToDollarState();
}

class _InrToDollarState extends State<InrToDollar> {
  //create controller for inr
  TextEditingController inrController = TextEditingController();
  //creating the varibale to store the dollar value
  double dollar = 0.0;
  @override
  Widget build(BuildContext context) {
    return Scaffold(
      appBar: AppBar(
        title: Text('Inr to Dollar Converter'),
      ),
      body: Container(
        //give some padding
        padding: EdgeInsets.all(20),
        //here we need to create
        //one text filed to take input form user
        //and one button
        //so we will use here column widget
        child: Column(
          children: [
            TextField(
              controller: inrController,
              //some box around text field
              decoration: const InputDecoration(
                border: OutlineInputBorder(),
                //labele text
                labelText: 'Enter Inr',
              ),
            ),
            //need some space between
            //textfield and button
            const SizedBox(
              height: 30,
            ),
            //create convert button
            ElevatedButton(
              onPressed: () {
                //on pressing the convert we
                //should convert the inr to dollar
                setState(() {
                  dollar = double.parse(inrController.text) * 0.012;
                });
              },
              child: Text("Convert"),
            ),
            //for some space between button and text
            SizedBox(
              height: 30,
            ),
            //now we have to show dollar on screen
            //so we will use here text widget
            Text(
              "Dollar : " + dollar.toString(),
              //increse some font size
              style: TextStyle(
                fontSize: 30,
              ),
            ),
          ],
        ),
      ),
    );
  }
} dollar_to_inr.dart import 'package:flutter/src/widgets/framework.dart';
import 'package:flutter/src/widgets/placeholder.dart';
import 'package:flutter/material.dart';

class DollarToINR extends StatefulWidget {
  const DollarToINR({super.key});

  @override
  State<DollarToINR> createState() => _DollarToINRState();
}

class _DollarToINRState extends State<DollarToINR> {
  //create controller for dollar
  TextEditingController dollarController = TextEditingController();
  //creating the varibale to store the inr value
  double inr = 0.0;
  @override
  Widget build(BuildContext context) {
    return Scaffold(
      appBar: AppBar(
        title: Text('Dollar to inr Converter'),
      ),
      body: Container(
        //give some padding
        padding: EdgeInsets.all(20),
        //here we need to create
        //one text filed to take input form user
        //and one button
        //so we will use here column widget
        child: Column(
          children: [
            TextField(
              controller: dollarController,
              //some box around text field
              decoration: const InputDecoration(
                border: OutlineInputBorder(),
                //labele text
                labelText: 'Enter Dollar',
              ),
            ),
            //need some space between
            //textfield and button
            const SizedBox(
              height: 30,
            ),
            //create convert button
            ElevatedButton(
              onPressed: () {
                //on pressing the convert we
                //should convert the inr to dollar
                setState(() {
                  inr = double.parse(dollarController.text) * 82.40;
                });
              },
              child: Text("Convert"),
            ),
            //for some space between button and text
            SizedBox(
              height: 30,
            ),
            //now we have to show dollar on screen
            //so we will use here text widget
            Text(
              "Inr : " + inr.toString(),
              //increse some font size
              style: TextStyle(
                fontSize: 30,
              ),
            ),
          ],
        ),
      ),
    );
  }
} Step-by-step process https://youtu.be/705GO3kj9FU?embedable=true Also published here. The lead image for this article was generated by HackerNoon's AI Image Generator via the prompt "currency converter".

Building a Currency Converter in Flutter

In this article, I’m going to share some of the productivity hacks that save time in my day-to-day activities as a product manager. Hope you’ll find it useful. Here is what I will cover in this article: Creating and editing images like interface mockups and screenshots Editing text faster with hotkeys and special symbols Using a clipboard manager to save hours a day Building interface mockups to illustrate ideas quickly Other useful apps and macOS settings Top 3 ChatGPT plugin use cases Creating and Editing Images As a product manager, you will often find yourself having to support your communication with visuals like screenshots, roadmaps and mockups. Screenshot Hotkeys Here are the standard screenshot hotkeys in macOS. My personal favourite is since it doesn’t pollute your desktop with infinite files. Copy picture of selected area to the clipboard After activating this screenshot mode with + + + , you can take a full window screenshot by pressing the . Since hitting four keys at once is hardly perfect, you can assign your own keystroke combos. ^ ⇧ ⌘ 4 Spacebar The mouse pointer will switch to a camera icon 📷 and let you grab the full window - just like in the example above. Fast Image Annotation Sometimes taking screenshots is not enough — you need to draw attention to some part of the image, add text, or both. On macOS, you can use the standard app to annotate screenshots. Open a screenshot or paste one from your clipboard (by opening the Preview app and pressing + ). Preview ⌘ N Then you can start adding elements to your screenshots: + + to add a Text block for comments; ⌘ ^ T + + to add an Arrow; ⌘ ^ A + + and + + to add an Oval or a Rectangle. ⌘ ^ O ⌘ ^ R Editing Text Faster Be it in Slack, Figma, Google Workspace or plain old email, you are required to work with texts, so incorporating hotkeys into your text workflow will speed things up. Here are the ones that I use the most: + or will move the cursor to the beginning or the end of the line respectively; ⌘ left arrow right arrow + or will move the cursor word by word; ⌥ left arrow right arrow Use the key with the hotkeys from above to delete full words or lines in one keystroke; Delete To select parts of the text, use the same hotkeys and hold down . Shift Now that you can quickly navigate through text, here are some quick actions that work in most apps: + marks the selection in bold; ⌘ B + underlines the selection; ⌘ U + the selection; ⌘ I italicises + turns the highlighted text into a clickable link; ⌘ K + + + is particularly useful in Google Workspace: it inserts the clipboard text and inherits the target formatting, and comes in handy when you copy from a document with one text font to a document with a different font. ⌘ ⌥ ⇧ V Ever wanted to insert emojis directly into text? The + + combo opens the macOS emoji and special character picker. ⌘ ^ Space These are some basic clipboard tricks. But if you’re looking for something truly special to boost your clipboard productivity, look no further than the Paste app. Using the Clipboard Manager Paste Your clipboard only remembers the last item you copied by default, which is particularly annoying when you need to frequently copy and paste, like when writing an article about hotkeys and keeping special characters at hand. When you copy something new, the previously copied item vanishes and makes way for the new item. To get it back, you need to find it and press + once again. ⌘ C This is when the Paste app comes in handy — it stores all the things you copied in a visual interface so that you can select what exactly you need to copy right now. There are plenty of other clipboard managers for both macOS and Windows, like Ditto and CopyLess. I prefer Paste because of its visual interface and cross-platform capabilities. Building Interface Mockups When talking to engineers or sharing design ideas with the team, it's important to illustrate ideas with visuals. Figma Even if your design team is not using or you don’t have edit access to any of the design projects, I strongly recommend installing the Figma desktop app and checking its to familiarise yourself with the key concepts and elements. Figma beginner tutorials This way, you can create your own space to jiggle with wireframes (e.g. to see if certain text fits on a single line). If you are up for more digging or need a feature explained, Figma has an amazing where you can explore projects that others have published (like an ) or basic UI elements and states (e.g. ). community e-commerce mobile app iOS system alerts Finally, with the Figma mobile you can test how a design feels on a real device: , mirror mode Log in to your Figma account on a mobile device; On the laptop, select a frame that you want to test, and it will automatically appear on your mobile device’s screen; Voila! The controls will remain static, but scrolling will work and you’ll be able to see how the design looks on a mobile phone. Balsamiq If you don’t need pixel-perfect interfaces but rather a low-fidelity sketch, I recommend checking out , available on the desktop and in the browser. For me, it came in handy while creating single-page interface mockups from scratch (like feature configuration pages). Balsamiq Other Productivity Tips and Apps for converting videos to GIFs. It works fast, has a clean interface, and comes in handy when inserting screen recordings into Google Slides. Gifski is a simple, yet powerful to-do list that can be infinitely nested and zoomed in and out. I use it for tracking my personal tasks during the week. It works across platforms with blazing speed. Workflowy can move and resize apps and browser windows with hotkeys. It is extremely powerful (e.g. you force an app to take up 1/3, 1/4, 1/5, or 1/6 of the screen), and can be overkill for some, but even using 2 or 3 simple hotkeys helps me arrange my screen for multitasking quickly Rectangle I use as a backlog of articles and documents to read. When it syncs on your phone, it downloads all the pages’ contents, which is incredibly useful when you’re commuting by metro or flying with no internet access. Pocket is the most lightweight browser extension that visualises all your bookmarks for easy navigation. Bookmarks Cards ChatGPT Plus Use Cases The free version of ChatGPT needs no introduction already, but here are my top 3 use cases for the Plus version that you may not be aware of: Retrieving industry benchmarks The ChatGPT browsing plugin does a lot of work for me: it provides a concise summary of what it was able to retrieve with references that you can open and double-check in case it is hallucinating. Without this feature, I would look up, for instance, the average CTR of a Facebook e-commerce banner in the US, then go through dozens of SEO-optimised pages written by ad platform startups and see that their contents are largely the same. Creating sequence diagrams With the Show Me plugin, whatever ChatGPT can output as text can now be converted into a diagram, and pre-generated diagrams can be edited either by asking ChatGPT or by hand (you can go directly to the Show Me website and edit). Retrieving documentation When I’m using SQL, Amplitude, Looker or any other day-to-day analytical tools, ChatGPT with a browser plugin can explain how to use certain functions of these tools. Unlike reading the documentation yourself, you can now get answers exactly the way that you phrased your question. Furthermore, it provides relevant auto-generated examples so that you can see precisely what to do. Thank you for sticking around! I hope you will find some of these productivity tips and tricks useful in your line of work. Let me know which ones you liked the most, and what topics you’d like to hear more about. If you have any questions, feel free to reach out . here

Productivity Hacks for Busy Product Managers: Useful Apps and Hotkeys

Do you know you can build flutter apps in Python?😮 Flutter is quite popular in the software development world. Let’s deep dive into the world of building flutter apps with Python!🙂 About FLET Before we begin, what is FLET? FLET enables developers to easily build real-time apps in Python. web, mobile, and desktop The crazy thing is that no front-end experience is needed, and although the mobile version is still in development, we can still rely on the Progressive Web App. Mind-blowing features of FLET It is powered by flutter. You can bring an app to life in a few minutes. It has a simple architecture. Apart from Python, other languages like Go, C#, etc. will also be supported. Amazing, right??🙂🔥 How to Install FLET To install FLET, you use this command: pip install flet Remember: This is for people using a Python version less than version 3. Otherwise, pip3 install flet To upgrade your pip to the latest version. pip install -- upgrade pip 🏊‍♀️ A brief dive into Flutter The UI toolkit , created by , lets programmers create apps with top-notch user interfaces. In contrast to Java itself, Flutter employs Dart programming, an object-oriented language that is much simpler to learn. Flutter Google For the new Flutter version(3.4.0-34.1.pre), we can build apps for with a single code-base. mobile (Android/IOS), MacOS, Web, Linux, and desktop Note: When using multiple operating systems, we obviously need to make certain adjustments. Building a very simple application with FLET Let’s build a simple application with FLET. For example, a Counter application that has One Text-field Two buttons for Increment & Decrement Code Snippets First, we need to import and other features essential for the counter App (e.g. widgets). FLET import fletfrom flet import Row, icons, IconButton, TextField, Page; We define the main function which is the root of our application and also set a title. Let’s arrange the widgets because the base part of the app is already ready. To add a widget, we use the page.add(widget…..). def main(page: Page):
page.title = "Counter App"
page.vertical_alignment = "center" For the rows of the app, we would consider the… Text Field and Buttons page.add(Row([
   IconButton(icons.REMOVE),
   TextField(text_align="center",value="0", width=100)
   IconButton(icons.ADD)
  ],
 alignment="center")
) Now, We define two functions that will handle the press events. //Decrement
def minus(e) :
tf.value = int(tf.value) -1
page.update()//Increment
def plus(e) :
tf.value = int(tf.value) + 1
page.update() refers to Text-field we placed into the Row. The next thing, we used to the text-field value we have to wrap that within a variable itself. ‘tf’ This is the overall code snippet: import flet
from flet import Row, icons, IconButton, TextField, Page;

def main(page: Page):
  page.title = "Counter App"
  page.vertical_alignment = "center"
  tf =TextField(text_align="center",value="0", width=100)

#Functions
  def minus(e):
    tf.value = int(tf.value) -1
    page.update()

  def plus(e):
    tf.value = int(tf.value) + 1
    page.update()

#Widgets
  page.add(
    Row([
       IconButton(icons.REMOVE, on_click=minus),
       tf,
       IconButton(icons.ADD, on_click=plus)
    ],
    alignment="center")
)
flet.app(target=main,view=flet.WEB_BROWSER) The final part is to run the application. flet.app(target=main) We add this command which directly targets the “main” To run the app, type this command python filename.py Note that you have to be in the same directory as the file. The UI: As I stated in the section, it can also be used on your web browser. A brief dive into Flutter A small adjustment is needed for that; simply add flet.app(target=main,view=flet.WEB_BROWSER) Here is the final result https://youtu.be/dN6auCBICds?embedable=true You can learn more about FLET here: https://flet.dev/docs/roadmap

A Brief Intro to FLET: Building Flutter Apps with Python

Una breve introducción a FLET: creación de aplicaciones de Flutter con Python

Une brève introduction à FLET : créer des applications Flutter avec Python

FLET का संक्षिप्त परिचय: Python के साथ फ़्लटर ऐप्स बनाना

FLET の簡単な紹介: Python を使用した Flutter アプリの構築

Uma breve introdução ao FLET: criando aplicativos Flutter com Python

Giới thiệu ngắn gọn về FLET: Xây dựng ứng dụng Flutter với Python

FLET 简介：使用 Python 构建 Flutter 应用程序

In this article, I want to share with you, how to create your python wrapper, that solves the basic problem of the tesseract engine – the small speed of recognizing multiple pages in one document. The basic idea is to use python’s built-in multiprocessing features to split documents into separate pages and run multiple tesseract engine instances for parallel page recognition. Tesseract uses one core to recognize images, in average cases, it will be enough, but if you have “heavy” documents, that have many sheets, it will be very slow, because tesseract by default will use only one CPU core. Why do we need tesseract? So, in some cases, when you cannot copy content from pdf-document (pdf has different formats itself, one of them – being a scanned pdf document, which is just an image of the actual text), we need software, that will recognize text from image. This technology is called OCR (Optical Character Recognition). One of the most popular and free OCR software is tesseract. Tesseract originally was developed by Hewlett-Packard in 2006, then it was sold to Google. Google made tesseract free and open source. Today, the tesseract is being developed by a group of enthusiasts for free. Set up the environment. For developing our python wrapper, we need the latest version of python, currently, it is 3.11, download link: https://www.python.org/downloads/release/python-3110/ and (for the virtual environment).  When you finished downloading python, run it on your computer. Don’t forget to select “Add python.exe to PATH” for running it from your PowerShell or . pipenv cmd.exe Next, install through the command in cmd.exe. Finally, you will see “Successfully installed”. Also, we need to install ( ) and for it. pipenv pip install pipenv pdf2image pip install pdf2image download a poppler You will see the “Successfully created virtual environment” and its full path. Remember the full path, you will use it in VS code IDE. Open VS Code and select python interpreter with the combination then select “Python: Select Interpreter”. ctrl + shift + p Select our newly created virtual environment from the menu. Next, we need to (v5.2.0.20220712) and put it inside our project folder (create a tesseract folder inside the project folder. download the tesseract engine After installation, we are ready to go. Programming. Create file and paste this code. main.py import tempfile
from pdf2image import convert_from_path
import pytesseract
import time


pytesseract.pytesseract.tesseract_cmd=r'C:/Users/Nuriq/Desktop/python_wrapper/tesseract/tesseract.exe'


def start_ocr(pdf_path, poppler_path):
    full_raw_text = ""

    with tempfile.TemporaryDirectory() as path:
        images_from_path = convert_from_path(
            pdf_path=pdf_path, 
            output_folder=path, 
            paths_only=True, 
            fmt="jpeg", 
            poppler_path=poppler_path, 
            dpi=250, 
            grayscale=True
            )
        for img_path in images_from_path:
            full_raw_text += pytesseract.image_to_string(img_path)
    return full_raw_text


if __name__ == '__main__':
    start_time = time.time()
    full_raw_text = start_ocr(
        'C:/Users/Nuriq/Desktop/python_wrapper/PublicWaterMassMailing.pdf', 
        'C:/Users/Nuriq/Desktop/python_wrapper/poppler-0.68.0/bin'
        )
    print(full_raw_text)
    end_time = time.time()
    print(f"it took: {end_time-start_time}") The line is the full path to our tesseract.exe engine. We use python “TemporaryDirectory” to store temporary files, which works faster than physically storing them on our HDD/SDD. To start using tesseract, we need to convert a single pdf document to images (tiff/jpeg files), where 1 page = 1 image. pytesseract.pytesseract.tesseract_cmd Store it in the top folder and run , where is a full path to the image. full_raw_text += by tesseract.image_to_string(img_path) img_path Finally, we got a 51-second result to proceed with all 32 pages ( took 8 seconds to convert pdf to images). Then write the next code. pdf2image import tempfile
from pdf2image import convert_from_path
import pytesseract
import os
import time
import concurrent.futures
from concurrent.futures import ProcessPoolExecutor


pytesseract.pytesseract.tesseract_cmd = r'C:/Users/Nuriq/Desktop/python_wrapper/tesseract/tesseract.exe'


def start_ocr(pdf_path, poppler_path):
    images_from_path = []
    with tempfile.TemporaryDirectory() as path:
        images_from_path = convert_from_path(
            pdf_path=pdf_path, 
            output_folder=path, 
            paths_only=True, 
            fmt="jpeg", 
            poppler_path=poppler_path, 
            dpi=250, 
            grayscale=True
            )
        with ProcessPoolExecutor(max_workers=os.cpu_count()) as executor:
            tasks = {executor.submit(pytesseract.image_to_string, img_path): img_path for img_path in images_from_path}
            for future in concurrent.futures.as_completed(tasks):
                page_number = tasks[future]
                data = future.result(), page_number[-5]
                yield data


def sort_text(text):
    return(sorted(text, key = lambda x: x[1])) 


def pdf_to_string(pdf_path, poppler_path):
    full_raw_text = start_ocr(
        pdf_path, 
        poppler_path
        )
    full_text = ""
    text = sort_text(full_raw_text)
    for page_text, _ in text:
        full_text += page_text
    return full_text


if __name__ == '__main__':
    start_time = time.time()
    text = pdf_to_string(
        'C:/Users/Nuriq/Desktop/python_wrapper/PublicWaterMassMailing.pdf', 
        'C:/Users/Nuriq/Desktop/python_wrapper/poppler-0.68.0/bin'
        )
    print(text)
    end_time = time.time()
    print(f"it took: {end_time-start_time}") Run it and finally, we got 7 seconds (on Core i7 12700), which is than the previous code sample. 7.3x faster Conclusion. Finally, using we run 19 instances of tesseract for every file on the list and do parallel calculations. Then we save every process result, and we put the page number to it. After all, using the function, we sort them and merge them into one text. ProcessPoolExecutor sort_text()

Creating a Wrapper for Tesseract is Several Times Faster Than PyTesseract

Hi! My name is Oleksii Avilov. I work as an ML engineer at Ukrainian startup ZibraAI. I am engaged in research, automation, and the development of ML solutions. I mainly work on image-to-3D generation tasks, and now also text-to-image. Many 3D artists and illustrators believe that artificial intelligence will eventually replace them. I look at this challenge differently. At ZibraAI, we develop solutions to simplify the work of game designers using artificial neural networks. We don’t want to replace people, but strive to help them by making the most time-consuming and boring processes faster. For the past few years, I have been fascinated by generative art – the art generated with the help of artificial intelligence. At the beginning of last year, we worked on 3D model generation and tested different approaches, including text-to-image generation. It started rather as a hobby, and it probably would have remained a hobby, if not for the war. In this blog, I want to cover different approaches to generating images using artificial intelligence and share the story of how we used generative art to draw attention to the war in Ukraine. The dawn of generating images from text The first attempts to generate images from text began in the mid-2010s, with the appearance of Generative Adversarial Networks (GANs). Generative Adversarial Network is a system of two artificial neural networks that compete with each other. One network (the generator) generates images based on textual descriptions, while another (the discriminator) evaluates them. During the training phase, the goal of the generator is to trick the discriminator by creating a synthesized image as similar to the real one as possible. The task of the discriminator is to accurately distinguish real images from synthesized ones. Here is an example of such a generation created in 2016: It was a then, although it certainly looks pretty bad now breakthrough Later several alternative algorithms appeared that generated images from text queries, but there was no visible improvement in terms of quality compared to GANs. A good overview of image generation from long ago to today can be found here on . GitHub The next stage of development – DALL·E and CLIP Only at the beginning of last year, there came major changes in the field of text-to-image. OpenAi introduced two solutions which, in my opinion, gave the start to the revolution in image generation that we observe now. They were and neural networks. DALL·E CLIP DALL·E neural network is based on GPT-3, the third generation of natural language processing algorithm from OpenAi, and it has a transformer architecture that extends text sequences with special image tokens, which are then transformed into images by another model (decoder). Compared to previous solutions, DALL·E showed great   progress in terms of synthesized images quality. The level of generalization which characterizes the neural network deserves special attention. Thanks to it, DALL·E can generate samples that it hasn’t seen during training (examples that are missing from the training dataset). For example, these avocado-shaped chairs have become OpenAI's signature. CLIP has got less hype upon release, but I believe it made a more significant contribution to the development of text-to-image generation. CLIP perfectly links images to text and consists of two encoders, one for text and one for images. [https://openai.com/blog/dall-e/> Unlike DALL·E, the trained CLIP weights have been made freely available by the developers. After that, many CLIP-based text-to-image solutions (like , , , and others) have been introduced. BigGAN+CLIP VQGAN+CLIP CLIP Guided Diffusion And since then, something crazy has been going on in this industry. There’s an ongoing competition between open source solutions from the community of independent developers (like , , ) and commercial models of large (some of them not) enterprises ( from OpenAI, from Google Research, , and others). More and more generative content appears in the world. Disco Diffusion Latent Diffusion Stable Diffusion DALL·E 2 Imagen Midjourney Midjourney. Source Midjourney. Source DALL·E 2 generations: Image sources in the carousel: , , , , , , , , , , , . [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] Imagen generations: Image sources in the carousel: , , , , , . [1] [2] [3] [4] [5] [6] Stable Diffusion generations: Image sources in the carousel: , , , , , , , , . [1] [2] [3] [4] [5] [6] [7] [8] [9] Here are some generated in Stable Diffusion: cool GIFs Although big corporations usually have access to more computing power, many generative artists prefer open source solutions. The advantage of such solutions is that everyone can participate in product development and try it, with the technologies themselves evolving and developing quickly. Moreover, the choice between open source products and commercial solutions is often due to censorship. For example, OpenAI prohibits generating images on sensitive topics and bans users after several attempts to generate them. The word "Ukraine" is also prohibited for text input. Censorship in DALL·E 2 Generation of images about the war in Ukraine In the first week of the full-scale war, our team decided to use our expertise to help raise funds to rebuild cities destroyed by Russia and to remind the world that the war was not over. That’s how the Sirens Gallery project was born. . Upon getting started we conducted research and chose the approach based on Disco Diffusion. At that time, version was in public access, and now has already been released. Disco Diffusion is based on the class-conditional with CLIP, which links text prompts with images (you can play with the artificial neural network ). 4.1 5.6 diffusion model from OpenAI here ** Day 1 — Ghost of Kyiv We've explored the parameters, available models, and art styles and started generating pictures (you can find a parameters guide , and Disco Diffusion styles guide ). To get a better image quality, a super-resolution model has been added to the pipeline. Furthermore, we made a convenient interface for internal needs. Now everything works in the Telegram bot. here here But in fact, the technical part was not the hardest one. The biggest challenge was to make the war timeline. In order to choose the most important events, and write stories about them, we had to review a lot of events and photos.  You can't help but take it to heart when you read about yet another atrocity of Russian soldiers, another story of raped women and children, or look at photos from Bucha or Izyum, where the streets are littered with the corpses of civilians. In total, Sirens Gallery consists of more than2,000 pictures generated by artificial intelligence based on textual descriptions of the most important war events. 1,991 paintings have been offered for sale as NFTs on the and platforms. At the moment, we have already sold part of the paintings with an overall cost of 250,000 UAH (~$7,000). We transferred these funds to three charity projects on platform (the report is available ). Opensea.io Paras.id Dobro.ua here Our goal was to draw the world community's attention to the horrors that Russia (the terrorist state) commits in Ukraine. At the same time, we wanted to show the courage, bravery, and humanity of Ukrainians and raise funds to help the victims of the war. ** Day 85 — Heroes of Azovstal hold defense of Mariupol for 85 days Ongoing technology development Technologies evolve constantly. Some solutions provide more realistic generations. Since we have started working on Sirens Gallery, Stable Diffusion, Imagen, Midjourney, and DALL·E 2 have been released. A few weeks ago, the developers . Thanks to this a lot of new solutions began to appear. For example, a with an interface common to Disco Diffusion users, where you can generate your own images, 2D and 3D animations either from text or from image. released the Stable Diffusion weights collab The Stable Diffusion team has also created an API. Based on it, some ), , and plugins have already appeared. There is an option with texture generation for and even a solution for open-source digital painting project. You can learn more about these plugins and . Photoshop GIMP Blender Blender Krita here here You can look through a large of Stable Diffusion generation. And there’s also a for newbies on how to get started with Stable Diffusion on your PC. database guide At Sirens Gallery we haven’t counted on technology perfection, our project is rather about what you can achieve using new technologies. You can find our stories and corresponding AI-generated illustrations on our social networks: and . They are also available on the Sirens Gallery project . Instagram Twitter website Day 4 — Russian helicopters get destroyed in Chornobayivka Recently, together with , we held the first offline exhibition of AI-generated pictures, which depict the stories of saving Ukrainian children from the war. In the following months, the paintings will be presented on three continents. Afterward, they will be sold in New York. All funds will be transferred to help Ukrainians who became victims of Russian aggression. Save Ukraine

Text-to-Image: How AI Illustrates the War in Ukraine and What You Need to Generate Your Own Art

Anyone who has encountered non-trivial bugs on android probably knows that, sometimes, it takes a lot of time and effort to fix them. Other times, the situation is even worse - it takes a lot of time to reproduce the bug. There are cases when there is nothing to do but restart the application under different conditions: with a network, without a network, without cache, with a bad network, with erroneous responses from the server, etc. The situation can be aggravated by the fact that constant restarts require a lot of long single-type actions, e.g. filling fields with data. In this article, I want to describe one possible solution to this problem. Let’s get to it Observations and Assumptions It is assumed that you are familiar with adb This solution is suitable for filling fields of XML, compose, webView, and other visual types as well as other applications. Running a script from a scratch file sometimes causes an error. A quick way around the problem is to create another . kotlin There is currently no clean solution. scratch file Selecting an example Let's say we have an application that consists of two activities: and . To go to , you need to fill in the fields in and press : LoginActivity MainActivity MainActivity LoginActivity Enter Code: package com.leonidivankin.draftandroid.articles.autofill

import android.os.Bundle
import androidx.appcompat.app.AppCompatActivity

class MainActivity : AppCompatActivity() {
    override fun onCreate(savedInstanceState: Bundle?) {
        super.onCreate(savedInstanceState)
        title = "MainActivity"
    }
} <?xml version="1.0" encoding="utf-8"?>
<androidx.constraintlayout.widget.ConstraintLayout xmlns:android="http://schemas.android.com/apk/res/android"
    xmlns:app="http://schemas.android.com/apk/res-auto"
    xmlns:tools="http://schemas.android.com/tools"
    android:layout_width="match_parent"
    android:layout_height="match_parent"
    android:padding="24dp"
    tools:context=".articles.autofill.LoginActivity">

    <EditText
        android:id="@+id/editTextEmail"
        android:layout_width="match_parent"
        android:layout_height="wrap_content"
        android:hint="login@gmail.com"
        app:layout_constraintBottom_toTopOf="@+id/editTextPhone"
        app:layout_constraintTop_toTopOf="parent"
        app:layout_constraintVertical_chainStyle="packed" />

    <EditText
        android:id="@+id/editTextPhone"
        android:layout_width="match_parent"
        android:layout_height="wrap_content"
        android:hint="+1 650 123 4567"
        app:layout_constraintBottom_toTopOf="@+id/editTextPassword"
        app:layout_constraintTop_toBottomOf="@+id/editTextEmail" />

    <EditText
        android:id="@+id/editTextPassword"
        android:layout_width="match_parent"
        android:layout_height="wrap_content"
        android:hint="123456"
        app:layout_constraintBottom_toTopOf="@+id/buttonEnter"
        app:layout_constraintTop_toBottomOf="@+id/editTextPhone" />

    <Button
        android:id="@+id/buttonEnter"
        android:layout_width="match_parent"
        android:layout_height="wrap_content"
        android:text="enter"
        app:layout_constraintBottom_toBottomOf="parent"
        app:layout_constraintTop_toBottomOf="@+id/editTextPassword" />

</androidx.constraintlayout.widget.ConstraintLayout> package com.leonidivankin.draftandroid.articles.autofill

import android.content.Intent
import android.os.Bundle
import androidx.appcompat.app.AppCompatActivity
import com.leonidivankin.draftandroid.databinding.ActivityLoginBinding

class LoginActivity : AppCompatActivity() {

    private lateinit var binding: ActivityLoginBinding

    override fun onCreate(savedInstanceState: Bundle?) {
        super.onCreate(savedInstanceState)
        binding = ActivityLoginBinding.inflate(layoutInflater)
        setContentView(binding.root)
        title = "LoginActivity"

        binding.buttonEnter.setOnClickListener {
            if (binding.editTextEmail.text.toString() == "login@gmail.com"
                && binding.editTextPhone.text.toString() == "+1 650 123 4567"
                && binding.editTextPassword.text.toString() == "123456"
            )
                startActivity(Intent(this, MainActivity::class.java))
        }
    }
} To move to , the fields must be filled in, and their values must meet certain requirements.
For example, take the following values: MainActivity email: phone: +1 650 123 4567 login@gmail.com https://youtu.be/oX3NIZMYpyo As you can see, it takes about 1 minute from startup to . This is still on the assumption that there were almost no errors in filling the data. Sometimes, finding complex bugs requires multiple rebuilds of the application and restarts on the device. This means constant filling of fields to get further into the app. MainActivity It turns out that with each restart we would lose about 1 minute extra. The ideal would be to jump directly into , bypassing , but in large projects, this is most often impossible due to the need to get a token from the server. MainActivity LoginActivity Another possible solution to this problem would be to assign default values to these fields. However, this is not always possible, because the fields may be in an external or in a . In this case, you won't be able to set default values. SDK webView As we continue I’ll show you how you can optimize the filling of the fields with kotlin-script. Autofill fields Finding the necessary adb commands I will use the commands to fill the text fields. To fill one text field, we need: adb Click on the text area. Fill it with text. To click on a text area, you must use the command: adb adb shell input tap x y where x and y are coordinates relative to the upper left corner of the device in pixels (px). They can be found in different ways: There is an that knows how to determine the click coordinates: adb-command adb shell getevent -l But there are difficulties with the translation of coordinates. You can read more . here On the real device, you can enable the developer's settings to display click coordinates or grid. You can add a listener, which will give absolute coordinates at each click. You can also find out through . Layout Inspector In this example, I am more comfortable using paragraph 4. Launch the Layout Inspector. Connect to the current process. Select the necessary . view Check its coordinates. It will depend on the size and type of the device. Most likely, your coordinates will be different. Note that these coordinates are in . And the command needs coordinates in . In my case it is x: 243=72, y: 3643=1092 dp adb px The final command turns out to be: adb shell input tap 72 1092 Go to the terminal and enter this command: The text input field is now active. To find out the pixel density, you need to go to : Device Manager In my case it is . Here the ratio between and is 3. The ratios are described in more detail . xxhdi dp px here Next, you must enter the desired text in the field. To do this, use the command: adb shell input text 'login@gmail.com' Enter it into the terminal and check that the necessary is entered.
Similar actions are carried out with other text fields: . Next, we need to click on the button, for this, there is a command: email phone, password adb shell input tap 72 1497 I've already put the coordinates I need there. We found all the necessary to fill in all the text fields and click on the button. Here is the complete list: adb commands adb shell input tap 72 1092
adb shell input text 'login@gmail.com'
adb shell input tap 72 1227
adb shell input text '+1 650 123 4567'
adb shell input tap 72 1362
adb shell input text 123456
adb shell input tap 72 1497 Now we need to build a script from these commands that will execute them instead of us. Creating a kotlin script You can build a script in different ways and in different languages. Here are some of the options: Make a . bash script Make a plugin for . Android Studio Make , etc. Gradle task But each of these methods has disadvantages. The bash script is difficult to maintain, the plugin for needs to be maintained for new versions of the environment, gradle tasks take a long time to run due to dependencies with and . Android Studio gradle gradle wrapper So I use another way - . The kotlin language is more familiar to the android developer, the script is independent of the project, android studio, and gradle. kotlin-script (kts) Create a scratch file using (Windows). In the window that appears, select : Ctrl+Alt+Shift+Insert Kotlin This will create a file. Notice the green arrow. This means that the file is executable and can be run. Uncheck so that the script won't restart every time you change the file. scratch.kts Interactive mode There is a special method that allows you to execute an from a kotlin script: adb command Runtime.getRuntime().exec() Insert all of our commands. My complete scratch file is as follows: fun run(){
   run("adb shell am force-stop com.leonidivankin.draftandroid")
   run("adb shell am start -n \"com.leonidivankin.draftandroid/com.leonidivankin.draftandroid.articles.autofill.LoginActivity\" -a android.intent.action.MAIN -c android.intent.category.LAUNCHER")
   Thread.sleep(2000)
   run("adb shell input tap 72 1092")
   run("adb shell input text 'login@gmail.com'")
   Thread.sleep(500)
   run("adb shell input tap 72 1227")
   run("adb shell input text '+1 650 123 4567'")
   Thread.sleep(500)
   run("adb shell input tap 72 1362")
   run("adb shell input text 123456")
   Thread.sleep(500)
   run("adb shell input tap 72 1497")
}

fun run(command: String){
   Runtime.getRuntime().exec(command)
} This file consists of the we defined above with some additions. adb commands Note that I inserted the lines first: fun run(){
   run("adb shell am force-stop com.leonidivankin.draftandroid")
   run("adb shell am start -n \"com.leonidivankin.draftandroid/com.leonidivankin.draftandroid.articles.autofill.LoginActivity\" -a android.intent.action.MAIN -c android.intent.category.LAUNCHER") These lines are needed when the application is running and you need to restart it. I also put the command in a separate method to reduce the amount of code: exec() fun run(command: String){
   Runtime.getRuntime().exec(command)
} Note that some commands have between them. This is to wait for the application to start or the previous command to be entered. The point is that adb commands are executed asynchronously. And there is a probability that the focus will be shifted to the next text field without waiting for text input in the current one. Thread.sleep() Total transition time to from launching is not more than 5 seconds, which reduced the time by 20 times. MainActivity https://youtube.com/shorts/R484DB_RScc?feature=share Conclusion As Steve McConnell wrote in Perfect Code, programming is a complicated thing. It gets even more complicated because you have to look for complex and nontrivial bugs. So, from my point of view, the role of any optimizations is rather high. Despite the fact that within the framework of the whole project, they seem microscopic. Try to optimize the launch of your application by autocomplete fields.
If you find bugs or know how else to improve the script, post in the comments.

Autofill Fields with a Kotlin Script

Language: An Introduction to the Study of Speech: Chapter 4 

Practices Used in eLearning Video-Content Protection

Ye, the rapper formerly known as Kanye West, has been the talk of social media because of his latest posts about Pete Davinson. His posts are simply entertaining for some people as they see a celebrity gone "crazy." For others, it's a call to action to protect their favorite celebrity. A third party considers this scandal to be a dangerous situation that can escalate to violence or that it's already a violent act as it exposes Kim and Pete. In this slogging thread, we discuss whether or not we, as a society, enable violent behavior just because it is entertaining and trending. This Slogging thread by Mónica Freitas, Sara Pinto and Linh Smooke occurred in slogging's official #pop-culture channel, and has been edited for readability. Mónica Freitas Feb 23, 2022, 11:59 AM What Ye’s online warfare reveals about the dangers of celebrity fandom https://www.nbcnews.com/tech/internet/ye-kanye-west-kim-kardashian-instagram-pete-davidson-tenbarge-goggin-rcna16858 Mónica Freitas Feb 23, 2022, 12:02 PM "Some fans of the rapper, formerly known as Kanye West, see his social media warfare as meme fodder, but others worry the attention is unhealthy." Mónica Freitas Feb 23, 2022, 12:02 PM "Ye posted the meme in the midst of a torrent of social media activity in recent days, in which the rapper and artist has professed his desire to get back together with Kardashian West and his anger toward Davidson. Ye, who has more than 14 million Instagram followers, has posted multiple creations made by fans and screenshots of comments left by fans. He has also posted screenshots of private text messages from Kardashian West and Davidson, including texts from his former wife asking him to stop posting their private conversations. In one text, Kardashian West cautioned Ye that Davidson was going to get hurt because of his posts. All the posts have since been deleted, but screenshots have spread across the internet." Mónica Freitas Feb 23, 2022, 12:03 PM " , and in some cases those groups have . In other instances, celebrities have utilized their fan bases to , sometimes also using the content their fans create. It’s a dynamic that can sweep up even well-meaning fans, while also fueling others who see no problem going to extremes to support their favorite celebrities." Plenty of celebrities have attracted devoted, and at times aggressive, online fan bases gone beyond fandom and spilled over into harassment silence critics or push their agendas Mónica Freitas Feb 23, 2022, 12:07 PM This recent buzz around Ye has received different reactions: for some, it's simply entertaining as they see a celebrity gone "crazy" for others, it's a call to action to protect their favorite celebrity a third party sees this scandal as a very dangerous situation that can escalate to violence or that it's already a violent act itself as it exposes Kim and Pete Mónica Freitas Feb 23, 2022, 12:10 PM My question here is: are we, as a society, enabling violent behavior just because it is entertaining and trending? Are platforms allowing way too much room for this kind of demeanor? Are platforms quicker to turn a blind eye when it comes to celebrities? 🔥 1 Sara Pinto Feb 23, 2022, 4:42 PM Mónica Freitas, I believe so, and worse, we are also enabling mental illness, in the sense that we laugh about it, and we're not helping others seek help. The internet is an open space, which is good, but when it comes to propagating behaviors like that, and there are no sanctions, it's concerning. Plus, this all aggravates when the fandoms, as you mentioned as well, support it. Some fans are so blind and put the celebrity in such a high place that they think everything they say is correct. 🔥 1 Mónica Freitas Feb 24, 2022, 6:24 PM Sara Pinto, you're entirely right. It's worrying how we can be so nonchalant about these things. Just now, as Russia invades Ukraine, I've seen people making jokes. I wonder if we're desensitized... and why? Sara Pinto Feb 28, 2022, 11:48 AM Mónica Freitas, whether or not we like it, social media has a lot of impact on the way we think. Even though there are informative people out there, there's also misinformation and careless actions. And we can't help but be influenced by it. Mónica Freitas Mar 1, 2022, 9:25 PM Sara Pinto, could it be that we tend to follow the masses? If the majority likes something, we tend to share the same opinion, and vice-versa? Sara Pinto Mar 1, 2022, 9:57 PM Mónica Freitas, I believe so, but for me, the worst part is that people follow them clueless about the impact of their actions. And I think we can't see it with this Ye case. People follow and support him, regardless of how it's affecting his wife's life. Mónica Freitas Mar 2, 2022, 4:06 PM Sara Pinto, I don't even think people think about Kim and their children. They're just laughing at Ye's overreaction or hating on Pete. Kim is probably terrified, but some people still see this as a man fighting for his wife... go figure. Linh Smooke Mar 2, 2022, 10:48 PM For anything Kardashian related - this is my go-to educational source!!! https://www.tiktok.com/@kardashian_kolloquium?lang=en 🔥 1 😂 1

Ye’s Online Warfare: Dangerous or Comical?

This Slogging thread by Ellen, Richard Kubina, Linh, Guy, Jeferson Borba, and Marcos Fabian occurred in HackerNoon's official #product channel, and has been edited for readability. Here at we have different departments. HackerNoon, Incredible and eccentric people are present throughout the various teams, but department roles differ. As previously announced in our product updates blog account, the marketing department is eager to regularly honour the work of our developers. This 'honours program' helps to keep everyone in the loop on the many cool things we're designing. As such, a discussion about the process was in order. Such a discussion usually requires a meeting. Below is a glimpse of how creative people try and schedule a meeting. Ellen, the makes an attempt to invite herself to a product meeting and is promptly met with suspicion (rightfully so). [Clearly marked satire.] Marketing Campaign Lead Read below: Ellen Feb 17, 2022, 3:02 AM Ellen invited self to meeting tomorrow, please be there ❤ richard-kubina Feb 17, 2022, 4:13 PM Sorry, I am going to be resting today, I feel really tired with a sore throat. Slept in by a couple of hours already, kinda weird for me. I've been working on the Algolia instant search page and next.js v12 cache changes mostly. ❤ 1 🙏 1 Linh Feb 17, 2022, 4:45 PM Looks like we will have to do async since David and I are stuck in traffic back from the immigration office and won't be back until 10.30, unless people want to meet half an hour later? If not, let's post our weekly update here:  @Jeferson Borba @Marcos Fabian @Guy @andemosa @Kaushal Joshi @Beni @Storm Ellen Feb 17, 2022, 4:45 PM Omg is everyone afraid I will sing, I won't sing. :savage-joy: 2 Ellen Feb 17, 2022, 4:46 PM I'll just record the tutorial, it's not that complicated. Or half hour is good too. 💚 1 👌 1 Guy Feb 17, 2022, 4:46 PM I’m currently doing nothing but getting credits for sitting in a lecture theatre… can meet in half an hour or do a writeup - TLDR: NFTs Jeferson Borba Feb 17, 2022, 4:46 PM it can be 30min later or any other time for me, I'm always around anyways :savage-joy: :savage-joy: 1 ❤ 1 Marcos Fabian Feb 17, 2022, 4:47 PM Yeah we could do it 30 min later. Marcos Fabian Feb 17, 2022, 4:47 PM Or just write up the updates. Linh Feb 17, 2022, 4:47 PM Sounds good let's do half an hour later then! :high-five: 4 Ellen Feb 17, 2022, 4:47 PM 🙂 yay! Marcos Fabian Feb 17, 2022, 4:48 PM Ellen will probably still sing lol :savage-joy: 3 Ellen Feb 17, 2022, 4:48 PM HE WON'T STOP BLASTING COUNTRY MUSIC Ellen Feb 17, 2022, 4:48 PM I have sent my landlord 3 texts now. One video. :savage-joy: 2 Ellen Feb 17, 2022, 4:49 PM One text solely complaining about the genre. Guy Feb 17, 2022, 5:24 PM Still the same link? Seeing as Richy is OOO :extra-thinking: 1 Ellen Feb 17, 2022, 5:25 PM There's a valid question. Linh Feb 17, 2022, 5:30 PM I’ll start. Ellen Feb 17, 2022, 5:30 PM Let's start at the very beginning, a very good place to start Linh Feb 17, 2022, 5:31 PM [Zoom Link] Ellen Feb 17, 2022, 5:31 PM I think I remember how that song came up now haha [Meeting happening for hours, CEO about to announce big announcement, the meeting ends for everyone.] Guy Feb 17, 2022, 7:03 PM YOU'RE JOKING Guy Feb 17, 2022, 7:03 PM THE BIG ANNOUNCEMENT!!! Linh Feb 17, 2022, 7:03 PM LOL Linh Feb 17, 2022, 7:03 PM I MAKE MARCOS THE HOST Ellen Feb 17, 2022, 7:03 PM OK it didn't end for only me. [New Link Sent] Linh Feb 17, 2022, 7:03 PM Jeferson Borba Feb 17, 2022, 7:03 PM It's just to create some drama. Linh Feb 17, 2022, 7:03 PM HERE’S the NEW LINK. Marcos Fabian Feb 17, 2022, 7:04 PM Wait what happened? Guy Feb 17, 2022, 7:04 PM I think you were host and left. Ellen Feb 17, 2022, 7:04 PM Everything went dark. Marcos Fabian Feb 17, 2022, 7:05 PM Dammmmn sorry guys. The meeting was back on track in a jiffy and an exciting announcement was made. Stay tuned for that. The meeting raged on for a solid 2+ hours, with lots of notable information being discussed for the bettering of the platform and processes. You too can join the Slogging revolution here. Questions about product etc can be directed to this form.

How to Successfully Invite Yourself to a Meeting That Was Unsuccessfully Rescheduled

Cómo invitarse con éxito a una reunión que se reprogramó sin éxito

असफल रूप से पुनर्निर्धारित बैठक में खुद को सफलतापूर्वक कैसे आमंत्रित करें

Làm thế nào để tự mời thành công một cuộc họp đã được lên lịch không thành công

如何成功邀请自己参加未成功重新安排的会议

Stories, News & Mentions about ONETEXT

Onetext’s Stories on HackerNoon