AutoPR for paper information correction

.github/ISSUE_TEMPLATE/99-bulk-metadata-correction.yml

@@ -7,19 +7,16 @@ body:
   - type: markdown
     attributes:
       value: >
-        This form is activated by following a link from each paper page in the Anthology (e.g., https://preview.aclanthology.org/autopr/K17-1003/). The form will list the title, abstract, and authors in JSON format, which you can manipulate to make corrections, such as adjusting a title, correcting an author name, adding a missing author, or reordering names.
-
-        Please note to take care to preserve structure such as the `<fixed-case>` tag that is sometimes present in titles.
-
+        **This form is not meant to be used manually.** Instead, it is activated by following the "Fix metadata" link from each paper page in the Anthology (e.g., https://aclanthology.org/K17-1003/). Clicking this button displays a UI tool for modifying the title, abstract, and author list. Submission of that form will automatically populate the field below.
+  - type: markdown
+    attributes:
+      value: >
         Corrections will be processed in bulk on a weekly basis after verification by Anthology staff.
   - type: textarea
     id: data
-    attributes:
-      label: JSON code block
-      description: Please make corrections below, taking care to preserve the JSON structure (e.g., no trailing commas at the end of lists). If you add an author, do not worry about the ID unless you know what it is.
     validations:
       required: true
   - type: markdown
     attributes:
       value: |
-        **Note:** If you request an author name correction for yourself, please help ensure your name is correct for future publications by setting it correctly in Softconf/OpenReview.
+        **Important:** If you request an author name correction for yourself, please help ensure your name is correct for future publications by setting it correctly in Softconf/OpenReview.

.github/workflows/annotate-metadata-issue.yml

@@ -18,7 +18,6 @@ jobs:
             const hasRequiredLabels =
               labels.includes('correction') &&
               labels.includes('metadata');
-
             core.setOutput('has_required_labels', hasRequiredLabels.toString());
 
       - name: Parse JSON from issue body
@@ -56,9 +55,7 @@ jobs:
           script: |
             const anthology_id = core.getInput('anthology_id');
             const comment = `
-            Found ACL Anthology entry:
-
-            📄 Paper: https://aclanthology.org/${anthology_id}
+            Found ACL Anthology entry: https://aclanthology.org/${anthology_id}
 
             ![Thumbnail](https://aclanthology.org/thumb/${anthology_id}.jpg)
             `;

.github/workflows/reset-metadata-approval.yml

@@ -0,0 +1,38 @@
+name: Reset Approval on Edit
+
+on:
+  issues:
+    types:
+      - edited
+
+jobs:
+  reset-approval:
+    if: contains(join(github.event.issue.labels.*.name, ','), 'correction') || contains(join(github.event.issue.labels.*.name, ','), 'metadata')
+    runs-on: ubuntu-latest
+    steps:
+      - name: Leave a comment and reset approval status
+        uses: actions/github-script@v6
+        with:
+          script: |
+            const { issue, repository } = context.payload;
+            const owner = repository.owner.login;
+            const repo = repository.name;
+            const approvedLabel = "approved";
+
+            // Check if the issue has the "approved" label and remove it
+            if (issue.labels.some(label => label.name === approvedLabel)) {
+              await github.issues.removeLabel({
+                owner,
+                repo,
+                issue_number: issue.number,
+                name: approvedLabel
+              });
+
+              // Add a comment to notify about the edit
+              await github.issues.createComment({
+                owner,
+                repo,
+                issue_number: issue.number,
+                body: "Approval status has been reset after the issue was edited."
+              });
+            }

bin/anthology/utils.py

@@ -508,14 +508,25 @@ def parse_element(
     return attrib
 
 
-def make_simple_element(tag, text=None, attrib=None, parent=None, namespaces=None):
-    """Convenience function to create an LXML node"""
-    el = (
-        etree.Element(tag, nsmap=namespaces)
-        if parent is None
-        else etree.SubElement(parent, tag)
-    )
-    if text:
+def make_simple_element(
+    tag, text=None, attrib=None, parent=None, sibling=None, namespaces=None
+):
+    """Convenience function to create an LXML node.
+
+    :param tag: the tag name
+    :param text: the text content of the node
+    :param attrib: a dictionary of attributes
+    :param parent: the parent node
+    :param sibling: if provided and found, the new node will be inserted after this node
+    """
+    el = etree.Element(tag, nsmap=namespaces)
+    if parent is not None:
+        if sibling is not None:
+            parent.insert(parent.index(sibling) + 1, el)
+        else:
+            parent.append(el)
+
+    if text is not None:
         el.text = str(text)
     if attrib:
         for key, value in attrib.items():

bin/create_hugo_yaml.py

@@ -67,13 +67,15 @@ def export_anthology(anthology, outdir, clean=False, dryrun=False):
         if paper.parent_volume.ingest_date:
             data["ingest_date"] = paper.parent_volume.ingest_date
         data["title_html"] = paper.get_title("html")
+        data["title_raw"] = paper.get_title("xml")
         if "xml_title" in data:
             del data["xml_title"]
         if "xml_booktitle" in data:
             data["booktitle_html"] = paper.get_booktitle("html")
             del data["xml_booktitle"]
         if "xml_abstract" in data:
             data["abstract_html"] = paper.get_abstract("html")
+            data["abstract_raw"] = paper.get_abstract("xml")
             del data["xml_abstract"]
         if "xml_url" in data:
             del data["xml_url"]
@@ -138,6 +140,7 @@ def export_anthology(anthology, outdir, clean=False, dryrun=False):
         log.debug("export_anthology: processing volume '{}'".format(id_))
         data = volume.as_dict()
         data["title_html"] = volume.get_title("html")
+        data["title_raw"] = volume.get_title("xml")
         del data["xml_booktitle"]
         if "xml_abstract" in data:
             del data["xml_abstract"]