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
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
## 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
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