Fixes #28264. Streaming equi-join was including temporary columns used to materialize non elementwise join keys (`is_in`, `replace`) in the join payload. That caused wrong schemas, projection panics, and mismatches vs the in-memory engine.
## Checklist - [x] Appropriate tests were added - [ ] Any code changes were done in a way that does not break public API - [x] All documentation related to code changes were updated - [x] The new code follows the [contributor guidelines](ht
> [!warning] > **Draft / discussion only.** No real implementation yet — the only commit is a docstring stub that errors at call time. Stacked on top of #450; I'll rebase onto master once that merges. Hey @MartinuzziFrancesco — PR2 (#450) i
## Summary First PR of the **SciML Summer Fellowship 2026** track on ODE-based reservoirs (tracking issue: #397). Lands the core type hierarchy and a small dispatch refactor that lets continuous-time reservoirs plug into `collectstates` wit
## Summary Implements the second deliverable of the SciML 2026 Summer Fellowship (#397) — the `RCODEReservoirExt` package extension that makes `SciMLProblemReservoir` actually run. Builds on the core types from #446. - Continuous `_collects
Closes #1. ## What this fixes `roundEmail()` in `hikmat/public/game.html` rendered each Lesson Email slot's `label` and `labelHi` as raw template-literal interpolations into `innerHTML`. `tx()` is a passthrough language switch; it does not
Closes #5. ## What this fixes `hikmat.api.delete_student` is the right-to-erasure entry point for minors' data. It deletes the Student row and Lesson Attempt rows, but it has been silently leaving the corresponding Lesson Doubt rows behind.
## Description Fixes #221 To fix this, we can use a TypeVar to bound the input cls to VectorProtocol. This avoids the inconsistent inheritance issue. We also don't need the unreachable ignores anymore since the code now typechecks correctly
Resolves #4338. ## Summary Adds opt-in hybrid retrieval combining BM25 keyword scoring with semantic vector similarity. Backward compatible — existing workspaces are unaffected unless they explicitly switch the search mode. - **Strategy/orc
Fixes #6. ## What's going on `completeSession()` was making more PocketBase round trips than it needed to on the critical path between "जाँचो" and the celebration screen — and it scaled with class size. Two distinct issues on the same code
## What Replaces shorthand `<>` fragments with `<Fragment key={icon}>` inside the `.map()` call in the leaderboard page. ## Why Shorthand fragments (`<>`…`</>`) cannot accept a `key` prop. React requires a key on the top-level element retur
Closes #1. ## What Appends `content_progress` to the list of collections created by `scripts/setup_pocketbase.py`. Seven lines, no changes to anything else. ## Why `completeContent()` in `src/lib/db.ts` (introduced in efc11b3 — *\"feat: tra
## Summary Refs #2785. `OpenSearchVectorDBStorage.upsert()` previously issued one `helpers.async_bulk` call per invocation. The profile in #2785 shows the upstream pipeline calling `upsert()` 506 times for the demo workload, so the vector p
Closes #418. Implements Ceni, Pintore & Gallicchio, *Residual Reservoir Memory Networks*, IJCNN 2025 (doi:10.1109/IJCNN64981.2025.11227859), built on top of the RMN family introduced in #354. Following the family-naming convention discussed
## Summary Initial implementation of the `ReservoirComputingBenchmarks` library in `lib/` as discussed in #398. Implements the first three core benchmark tasks: - **Memory Capacity** (Jaeger 2002) — linear memory evaluation with per-delay a
Closes #2995. ## Summary `_call_protocol_parse_service` (used by `parse_docling` and `parse_mineru`) hardcoded the multipart field name as `"file"` and shipped Docling defaults that did not match the official [docling-serve API](https://doc
## Summary - Removes the `merge(st_after, (; :states => states_wo))` line introduced in c29cdad4 from `train!`. - Updates the `train!` docstring to point users to `return_states=true` / `collectstates` for accessing the feature matrix. ## M
Implement per-role LLM function separation as specified in #2741 Each role (extract, keyword, query) gets its own independently wrapped LLM function with a dedicated concurrency queue, enabling fine-grained control over model selection, con
Closes #417 Implements Ceni & Gallicchio (Neurocomputing 2024, doi:10.1016/j.neucom.2024.127966). Unlike ES2N which couples skip and nonlinear weights as (1-β) and β, ResESN uses independent α (skip) and β (nonlinear) scalars: h(t) = α·O·h(
References #5 - Implemented voice based banking with mock database and mock users. - Added test-cases and detailed documentation for setting up the project. - Added voice biometric authentication using Gaussian Mixture Models with MFCC feat
Pull Request resolves #45679 - [ x ] I read the [Generative AI policy](https://developer.joomla.org/generative-ai-policy.html) and my contribution is either not created with the help of AI or is compatible with the policy and GNU/GPL 2 or l
## Checklist - [x] Appropriate tests were added - [x] Any code changes were done in a way that does not break public API - [x] All documentation related to code changes were updated - [x] The new code follows the [contributor guidelines](ht
- Add `SVESM` model implementing the Support Vector Echo-State Machine from Shi & Han (IEEE TNN, 2007, https://doi.org/10.1109/TNN.2006.885113) - Widen `addreadout!` dispatch in `RCLIBSVMExt` from `ReservoirComputer` to `AbstractReservoirCo
## Summary This PR migrates `SemanticChunker` into `langchain-text-splitters` and aligns behavior for production use, including backward compatible initialization, deterministic threshold validation, and exact source span chunk assembly for
## Description Added page-Level tracking and citation. Now the references from the retrieval shows the exact page/range of pages from which the chunk was retrieved. <img width="451" height="71" alt="image" src="https://github.com/user-attac
fix #1 ### Changes 1. **Fixed test client tool names** (`tests/test_client.py`) - Changed `know_about_mospi_api` to `1_know_about_mospi_api` to match the server tool name - Replaced non-existent `lookup_mospi_codes` test with `2_get_indicat
closes #2 add .gitignore file to exclude unnecessary files and directories
fix #259 ## **PR Checklist | PR检查项** - [x] I have read and understood the [contributor guidelines](https://github.com/antgroup/agentUniverse/blob/master/CONTRIBUTING.md). | 我已阅读并理解[贡献者指南](https://github.com/antgroup/agentUniverse/blob/maste
Deleted the `gmail-mcp-server` and `mcp-headless-gmail` subprojects, and significantly revised the main README to reflect the new project structure, features, and user guidance for the AarogyaAI platform.
Included rules for ignoring Cursor IDE configuration files and directories to streamline project management and prevent unnecessary files from being tracked
Added intelligent chatbot endpoint for patient interactions, implemented chat history retrieval and clearing features, and updated frontend components to support chat history loading and display, improving user experience and interaction wi
Updated SYSTEM_NOTES and SYSTEM_PRESCRIPTION for improved clarity and structure, added error handling in AINotes component for better user feedback, and included User-Agent headers in various fetch requests across frontend components to imp
Implemented Medical Test Booking with recommended tests from the doctor and private booking
Improved transcriptions with better UI functionalities for viewing the same
Contributes to #3931 ## Describe changes I implemented/fixed comprehensive README restructuring to highlight pipeline serving capabilities and updated examples to achieve accurate representation of ZenML's dual capability for both batch and
Integrated AINotes to allow users to generate notes from transcripts, enhancing the note-taking functionality and providing an option to mirror AI notes into the scribe notes area.
Enhanced SpeechToText to manage session state and handle partial results. Updated Transcripts to display real-time transcription feedback and integrated session persistence for appointment transcriptions in DoctorPatientProfile.
Integrated Transcripts component and updated SpeechToText to handle live transcription sessions. Added fetching logic for appointment transcriptions and improved error handling for speech recognition support.
Added the scribe notes for the doctor and the appointment speech-to-text transcripts section to the patient profile in the doctor dashboard. Also made necessary schema changes in the DB to save the same and fetch and view on the frontend
Fix #3 Added appointments management for both patient and doctor
## Describe changes I implemented **a comprehensive README restructure** to achieve **better alignment with ZenML's pipeline serving capabilities and accurate representation of the updated examples**. The changes include: - **Added prominen
feat #4 Added Indian Markets for Analysis and Visualization
🔧 Core Components 1. AI Analysis Agent (src/agents/ai_analysis_agent.py) OpenAI Integration: Uses GPT-3.5-turbo for advanced market analysis Structured Data Classes: StockData and TechnicalIndicators for clean data handling Fallback System
closes #1682 feat: Implement SBOM Collection and Analysis **This PR introduces a new feature to OpenDigger for collecting and analyzing Software Bill of Materials (SBOM) data from top-level OpenRank projects.** This enhancement allows OpenD
Fixes #81 Added Top Loading Bar to both the Search and Send buttons SearchBar - Set progress to 30% on submit click Set progress to 100% after calling handleFormSubmit Twilio - Set progress to 30% on form submit Set progress to 60% during A
<!-- Thanks for contributing a pull request! Please ensure that your PR satisfies the checklist before submitting: http://scipy.github.io/devdocs/dev/contributor/development_workflow.html#checklist-before-submitting-a-pr Also, please name a
# Description Modified the __call__ method to accept **kwargs as a parameter. This will allow any additional keyword arguments to be passed to the method. You can then include these extra arguments when calling the interp method of the _xr_
Contributes to #80 - Added the "US Stock Market Data & Technical Indicators" dataset - Added Geometric Brownian Motion Model - Added Hidden Markov Model
Contributes to #539 This PR adds functionality to load and preprocess the MNIST dataset for handwritten digit classification using PyTorch. It includes the following key changes: Imports required libraries and modules for working with PyTor
Fixes #1 Changed the images in Readme.md for testing the ML Model/Platform. Changes, Improvements, Suggestions are appreciated for further design developements
Users can skip the re-installation of the build-time components if already present but still factor in the cases where it is necessary for them to reinstall for debugging purposes. The key aspects are: -Checking for existing Sundials and KL
Contributes to #62 I've worked upon this idea and I think we can implement the following: What is Curio: Welcome to Curio, a remarkable web application designed to address the language and communication barriers between mentors and learners
To #74 Updated the padding and font size for the footer to fit inside the mobile wireframe
Fixes #76 Added Github, LinkedIn and Twitter icons instead of Sign-Up, Sign-In and About
Fixes #44 Enabled Favicon by using the Csmall.png image file from src/assets/images
To #10563 Implemented a new function, read_image(), to read PNG images from file without OpenCV. RGB values for each pixel are stored in a nested list. Updated peak_signal_to_noise_ratio() to directly use the nested list instead of NumPy ar
Fixes #8865 * [ ] Add an algorithm? * [x] Fix a bug or typo in an existing algorithm? * [ ] Documentation change? ### Checklist: * [x] I have read [CONTRIBUTING.md](https://github.com/TheAlgorithms/Python/blob/master/CONTRIBUTING.md). * [x]
Fixes #9528 * [ ] Add an algorithm? * [x] Fix a bug or typo in an existing algorithm? * [ ] Documentation change? ### Checklist: * [x] I have read [CONTRIBUTING.md](https://github.com/TheAlgorithms/Python/blob/master/CONTRIBUTING.md). * [x]
Fixes #9598 The MLP class implements a simple feedforward neural network for classification. It has an input layer, a hidden layer with a tanh activation, and an output layer. The init() method initializes the weight matrices and bias vecto
# Description Users can skip the re-installation of the build-time components if already present but still factor in the cases where it is necessary for them to reinstall for debugging purposes. The key aspects are: -Checking for existing S
<!-- Thanks for contributing a pull request! Please ensure that your PR satisfies the checklist before submitting: http://scipy.github.io/devdocs/dev/contributor/development_workflow.html#checklist-before-submitting-a-pr Also, please name a
Fixes #3303 The key changes: - Use .get() instead of direct indexing to handle cases where the key is missing - Check if the value is a tuple using isinstance() - If it's a tuple, use the first index - Handle lithium plating the same way #
Fixes #9598 * [ ] Add an algorithm? * [x] Fix a bug or typo in an existing algorithm? * [ ] Documentation change? ### Checklist: * [x] I have read [CONTRIBUTING.md](https://github.com/TheAlgorithms/Python/blob/master/CONTRIBUTING.md). * [x]
### Describe your change: Fixes https://github.com/TheAlgorithms/Python/issues/9188 This implements random search with cross-validation for hyperparameter tuning in Python. The key differences from grid search CV are: - Instead of iterating
Fixes #9188 This implements random search with cross-validation for hyperparameter tuning in Python. The key differences from grid search CV are: - Instead of iterating through a grid of all combinations, it randomly samples hyperparameters
Fixes #21 Added a 20 second delay for users to register their next prompt. This will allow free users to add 3 requests per minute. - Used the time module of python - Added a function to check if an user is a free user - Added the rate limi