James Braza

Type error: `Value of type variable "_R" of function cannot be "AsyncClient"`

4

Upgrading to `pytest-asyncio==0.26.0`: ```python import pytest_asyncio from httpx import AsyncClient @pytest_asyncio.fixture(scope="function") def server_async_client() -> AsyncClient: ... ``` I get this error from `mypy`: ```none a.py:5:2: error: Value of type variable...

Request: rename `master` to `main`

2

**Is your feature request related to a problem? Please describe.** I repeatedly run `git checkout main` from muscle memory, and get denied because the main branch is `master`. Can we...

Minor reader bugfixes and fortified regression testing

1

This PR improves consistency across readers: - Tests equation parsing of Docling and PyMuPDF - Renames PyMuPDF's `image_dpi` arg to `dpi`, since (1) this parameter impacts tables too and (2)...

See PR title

Standardizing `system_prompt` usage across the repo

2

This PR fixes a bunch of minor issues with our prompts, removing thought burden: - Standardizes use of `system` prompt setting across the repo - Standardizes JSON-mode and non-JSON-mode `Context`...

`Docs.delete` doesn't remove from the `texts_index`

4

As of `paper-qa==5.29.1` I observed `Docs.delete` doesn't remove the associated `Text`s from the `texts_index`. This side effect is unexpected: a delete should not be just a partial delete. --- When...

Modernized and streamlined default LLMs

3

This PR: - Removes some one-off defaulted LLMs from the code base, keeping us centered on `Settings` - Updates LLM defaults to match the latest ~Anthropic~ OpenAI model - Went...

This PR: - Enables `DocDetails.doc_id` deduplication across files sharing the beginning (e.g. a main text and a supplemental information whose metadata is inferred to be identical) - Adds a `content_hash`...

Supporting Supplemental Information

3

### Background Sometimes a paper's supplemental information (SI) shows up in a separate document. The SI shares the same DOI as the main text. Ideally the title in the SI...

Due to https://github.com/Future-House/paper-qa/blob/v5.21.0/paperqa/docs.py#L409-L422, when building a local index, the `ParsingSettings.disable_doc_valid_check` is bypassed on: - Empty code files such as `py.typed` - Nearly-empty files such as `.python-version` Ideally we have a...