mc-providers 3.1.3

Search Providers package for Mediacloud

Media Cloud Providers Library

A package of search providers for Media Cloud, wrapping up interfaces for different social media platform.

Install with pip (pip install .) and the install.sh script.

Requires environment variables set for various interfaces to work correctly.

Releases > 4.0.0 will all be shared exclusively on github, and this project will be delisted from pypi before April 2025

Build

1. Bump the version number in pyproject.toml
2. Add a note about changes to the version history below
3. Commit the changes and tag it with a semantic version number
4. A github action will build and push the repository on committing a tagged version

Version History

• v3.1.3 - Accept "Seconds" argument to cache decorator. Update deployment action to track Minor releases
• v3.1.2 - Fix random sampling behavior in ES provider to be genuinely random, bugfix related to marginal sorting error, additional counters for fine-grained visibility
• v3.1.1 - Fix ES Provider to send None as last page pagination token
• v3.1.0 - Add new ProviderException classes to pass more meaningful errors to consumer processes
• v3.0.1 - Fix ES Provider to accept sort_{order,field} paging arguments like NSA-based Provider
• v3.0.0 - New "OnlineNewsMediaCloudProvider" using Elasticsearch DSL for direct access to the ES cluster. Retain old provider as "OnlineNewsMediaCloudOldProvider" for now.
• v2.2.0 - Added an optional argument to providers to toggle caching behavior, added more specific error on 504
• v2.1.1 - Bugfix
• v2.1.0 - Mediacloud news client code incorperated into this package
• v2.0.5 - Build-system in pyproject.toml
• v2.0.4 - reintroduce stopwords
• v2.0.3 - version bump for automatic releases
• v2.0.2 - respect domain filters on Media Cloud searches
• v2.0.1 - more work on caching strategies
• v2.0.0 - change CachingManager interface to support online news providers better
• v1.0.1 - fix default timeout option that applies across all providers
• v1.0.0 - Remove legacy Media Cloud, add timeout option to provider_for
• v0.5.3 - Temporary fix to onlinenews-mediacloud search handling
• v0.5.3 - Tweaks to onlinenews-mediacloud for compatibility with new database pattern
• v0.5.2 - Fix to allow override of chunk'ing in MC client
• v0.5.1 - Fix use of media cloud to respect domains clause on story list paging
• v0.5.0 - Integrate new mediacloud-news-client into onlinenews-mediacloud
• v0.4.0 - Specify custom base URLs via new string param to provider_by_name and provider_for
• v0.3.0 - Add support for paging through stories directly, and including text in paged results for speed
• v0.2.6 - Fixed querying by domain on new mediacloud system
• v0.2.5 - Alignment with new mediacloud system. Old onlinenews provider is now "onlinenews-mclegacy", "onlinenews-mediacloud" now queries the new index.
• v0.2.4 - Added support for api keys via "provider_by_name"
• v0.2.3 - removed support for API keys in environment variables- now expected as an argument in providers.provider_for
• v0.2.2 - transition to use the dedicated mediacloud-api-legacy package to avoid version conflictsgit
• v0.2.1 - add in a date hack to resolve a lower-level bug in the Media Cloud legacy count-over-time results
• v0.2.0 - add in support for Media Cloud legacy database
• v0.1.7 - corrected support for a "filters" kwarg in online_news
• v0.1.6 - Added support for a "filters" kwarg in online_news
• v0.1.5 - Added politeness wait to all chunked queries in twitter provider
• v0.1.4 - Added Query Chunking for large collections in the Twitter provider
• v0.1.3 - Added Query Chunking for large queries in the onlinenews provider
• v0.1.2 - Test Completeness
• v0.1.1 - Parity with web-search module, and language model
• v0.1.0 - Initial pypi upload