Repository navigation
Add 42 open-access papers across the seven thinnest topics (645 to 687) - #14
Merged
Merged
Conversation
The library gained no resources in the September maintenance pass. That
pass replaced 17 dead URLs and merged duplicate records; it added nothing.
This adds 42, and the seven topics chosen are the seven that were thinnest:
Mental Health & Wellbeing 12 -> 18
Energy & Just Transition 12 -> 18
Education 20 -> 28
Financial Inclusion 21 -> 27
Conflict & Humanitarian 24 -> 28
Social Protection 30 -> 33
WASH 29 -> 36
Total 645 -> 687.
Nothing here was written from memory. Every candidate went through three
independent checks and anything failing one was dropped:
1. Crossref must hold a record whose title matches the candidate at 0.85
token overlap or better. This is the check that matters most. An
earlier batch of 18 titles produced from recall rather than from a
search was rejected 17 times out of 18 — they were plausible, they
were not real, and without this step they would have entered the
library as fact.
2. Unpaywall must report an open-access location. One paper was dropped
here: the Kenya refugee regression-discontinuity study in the Journal
of Development Economics is not open access, and this is a library of
things a practitioner can open.
3. The URL must answer a live probe. A 403 is recorded as `blocked`
rather than discarded, which is this repository's own rule and the
reason its last audit did not delete a fifth of the collection.
Authors, year, journal, DOI and licence are taken from Crossref and
Unpaywall rather than typed. The descriptions are mine and say what the
paper found, including where the finding is inconvenient: the DRC WASH
trial that moved infrastructure and institutions and not one health
outcome; the four humanitarian cash studies reporting possible harm; the
off-grid solar review whose study-time result shows publication bias.
`type` uses this repository's three-value vocabulary (paper, book,
grey_literature) rather than the finer one the source data carries.
make build, make test (31 checks) and make contrast all pass.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01MrvR2NXsJFVRCeJZFCPuNL
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What does this PR do?
The September maintenance pass added no resources. It replaced 17 dead URLs and merged duplicate records, which was worth doing and was not what "new papers" means. This adds 42.
Library total 645 → 687.
Nothing here was written from memory
Every candidate passed three independent checks; anything failing one was dropped.
blockedrather than discarded — this repository's own rule, and the reason its last audit did not delete a fifth of the collection.Authors, year, journal, DOI and licence come from Crossref and Unpaywall rather than being typed. The descriptions are written here and say what each paper found, including where that is inconvenient: the DRC cluster-randomised WASH trial that moved infrastructure and institutions and not one health outcome; the humanitarian cash review where four of 23 studies report possible harm; the off-grid solar review whose study-time result shows publication bias.
typeuses this repository's three-value vocabulary (paper,book,grey_literature) rather than the finer one the sources carry.Type of change
Checklist
make build— 687 resources, all derived assets regeneratedmake test— 31 checks passmake contrast— passes🤖 Generated with Claude Code
https://claude.ai/code/session_01MrvR2NXsJFVRCeJZFCPuNL
Generated by Claude Code