Skip content under copyright

generate.py

@@ -62,6 +62,21 @@ def get_paras():
 
         root = tree.getroot()
 
+        # Skip content with copyright restrictions
+        # The copyright statements aren’t consistent.
+        # The ones that contain the text "public domain" (case insensitive)
+        # or nothing at all should be safe.
+        try:
+            rights = root.find(".//DC.Rights").text
+            rights = re.sub(r"\s+", " ", rights)
+        except:
+            rights = None
+
+        if rights and 'public domain' not in rights.lower():
+            print(f"Skipped {filename} due to copyright: {rights}", file=log, flush=True)
+            continue
+        continue
+
         for i, p in enumerate(root.findall(".//p")):
             xml = ET.tostring(p, encoding="unicode")
             text = get_text_content(p)
@@ -77,6 +92,7 @@ def get_paras():
                     "text": text,
                     "thml": xml,
                     "refs": refs,
+                    "rights": rights,
                     "all-MiniLM-L6-v2": emb,
                 }