Form submit issue

.github/scripts/fill_download_fields.py

@@ -38,7 +38,8 @@ def index_files(index_dir: Path | None, paths: list[Path]) -> list[Path]:
     the pull request changed.
     """
     if paths:
-        return [path for path in paths if path.name not in NON_DATASET_FILES]
+        # A pull request that removes an entry names a file that is gone.
+        return [path for path in paths if path.name not in NON_DATASET_FILES and path.exists()]
     if index_dir is None:
         return dataset_files()
     return sorted(p for p in index_dir.rglob("*.y*ml") if p.name not in NON_DATASET_FILES)

.github/workflows/fill_download_fields.yml

@@ -49,7 +49,7 @@ jobs:
       - name: Fill in the checksum and size
         run: |
           if [ "${{ github.event_name }}" = "pull_request" ]; then
-            files=$(git diff --name-only origin/${{ github.base_ref }}... -- 'emdatabase/index/*.yaml')
+            files=$(git diff --name-only --diff-filter=AM origin/${{ github.base_ref }}... -- 'emdatabase/index/*.yaml')
             if [ -z "$files" ]; then
               echo "no index file changed" | tee filled.md
               exit 0

docs/source/_build_docs.py

@@ -955,7 +955,7 @@ def generate_weights_html() -> str:
 
 
 def generate_add_dataset_html() -> str:
-    """The Add Dataset page: what a submission needs, and the two ways to send one."""
+    """The Add Dataset page: the issue form, and the CLI as the terminal route."""
     body = (
         '<main class="app-main"><div class="explainer">'
         '<div class="app-hero" style="padding:24px 0 6px">'
@@ -964,18 +964,6 @@ def generate_add_dataset_html() -> str:
         "an action turns it into the entry\u2019s YAML file, fills in whatever it can "
         "work out for itself, and opens the pull request.</p>"
         "</div>"
-        "<h2>What to have ready</h2>"
-        "<ul>"
-        "<li>A direct link to the file. A Google Drive share link or a Zenodo file "
-        "link both work.</li>"
-        "<li>The file name, if the link does not end in one.</li>"
-        "<li>A description of the data, the licence it is released under, and how it "
-        "was acquired \u2014 one technique or several.</li>"
-        "<li>The authors, one per line, as "
-        "<code>Name; Affiliation; ORCID</code>. The ORCID is optional.</li>"
-        "<li>For model weights: the class the checkpoint loads into, the framework it "
-        "was saved with, and the quantem versions it loads under.</li>"
-        "</ul>"
         '<p class="note">The checksum and the size are filled in automatically by '
         "downloading the file, so both can be left blank.</p>"
         '<div class="submit-row">'

docs/source/contributing.rst

@@ -183,6 +183,17 @@ that record, and moves ``latest`` to it. The file it looks for in the new
 record is the one whose name matches the current link; if the name has changed
 and the record holds more than one file, the run fails rather than guess.
 
+Removing an entry
+-----------------
+
+Delete the entry's file from ``emdatabase/index/``, run
+``python -m emdatabase._create_stubs`` so ``emdatabase/data/__init__.pyi`` no
+longer lists it, and open a pull request. Copies people have already downloaded
+are not touched. If the weekly weights job archived versions of a weights family
+to the ``weights-archive`` release, delete those assets too::
+
+    gh release delete-asset weights-archive <Family>_<date>.pt
+
 Hosting
 -------
 

emdatabase/data/__init__.pyi

@@ -326,20 +326,6 @@ class TutorialUNet(DownloadableDataset):
     Model weights hosted at https://drive.google.com; see ``.versions`` for the dated snapshots.
 
 
-    """
-    ...
-
-class TutorialUNet_instance(DownloadableDataset):
-    """
-    TutorialUNet_instance
-
-    Weights of the instance segmenting model from `neural_networks_02_unet.ipynb` available in the `quantem-tutorials` repository.
-
-    License: CC-BY-4.0
-
-    Model weights hosted at https://drive.google.com; see ``.versions`` for the dated snapshots.
-
-
     """
     ...
 
@@ -358,4 +344,4 @@ class ZrNbPrecipitate(DownloadableDataset):
     """
     ...
 
-__all__ = ['AlNanocrystals', 'AmorphousFilm4nm4DSTEM', 'ApoferritinApollo15eps', 'BilayerWS2', 'CuZnEELSMapping', 'CuZnHAADF', 'FeAlStripes', 'HREBSDStrainPatterns', 'InSituElectrochemGrowth', 'LSMOLineScan', 'LSMOLineScanLowLoss', 'LSMOSTOLineScan', 'LSMOSTOLineScanLowLoss', 'LayeredCuNb4DSTEM', 'MgONanoCrystals', 'NiEBSDLarge', 'PdCuSiCrystallization', 'PdNiPGlass', 'PeakDetectionPolymers', 'SPEDAg', 'TutorialUNet', 'TutorialUNet_instance', 'ZrNbPrecipitate']
+__all__ = ['AlNanocrystals', 'AmorphousFilm4nm4DSTEM', 'ApoferritinApollo15eps', 'BilayerWS2', 'CuZnEELSMapping', 'CuZnHAADF', 'FeAlStripes', 'HREBSDStrainPatterns', 'InSituElectrochemGrowth', 'LSMOLineScan', 'LSMOLineScanLowLoss', 'LSMOSTOLineScan', 'LSMOSTOLineScanLowLoss', 'LayeredCuNb4DSTEM', 'MgONanoCrystals', 'NiEBSDLarge', 'PdCuSiCrystallization', 'PdNiPGlass', 'PeakDetectionPolymers', 'SPEDAg', 'TutorialUNet', 'ZrNbPrecipitate']

emdatabase/index/PeakDetectionPolymers.yaml

@@ -31,7 +31,7 @@ PeakDetectionPolymers:
       affiliation: Stanford University
   kind: weights
   model:
-    class: other
+    class: quantem.core.ml.CNN2d
     framework: torch
   latest:
     url: https://zenodo.org/records/22311217/files/best.pth

emdatabase/tests/test_fill_download_fields.py

@@ -140,3 +140,15 @@ def test_an_entry_that_will_not_validate_is_not_written(script, index, tmp_path)
     assert code == 1
     assert path.read_text(encoding="utf-8") == before
     assert "Tomograhy" in summary
+
+
+def test_a_file_the_pull_request_deleted_is_skipped(script, index, tmp_path):
+    """A removed entry is named by the diff but is not there to read."""
+    _, directory, write = index
+    path = write()
+    gone = directory / "Removed.yaml"
+
+    code, _ = _run(script, directory, tmp_path, str(path), str(gone))
+    assert code == 0
+    assert "checksum" in path.read_text(encoding="utf-8")
+    assert not gone.exists()

emdatabase/tests/test_issue_route.py

@@ -438,7 +438,7 @@ def test_add_dataset_page_points_at_the_issue_form_and_the_cli(build_docs):
     assert build_docs._ISSUE_URL.endswith("/issues/new?template=new_dataset.yaml")
     assert "python -m emdatabase.new_dataset" in html
     assert "--kind weights" in html
-    assert "Name; Affiliation; ORCID" in html
+    assert "filled in automatically" in html
     for gone in ("<form", "<input", "spark-md5", "emdbBuildYaml"):
         assert gone not in html, gone