html: syntax fixes; resolve relative URLs; extract more XML fulltext URLs

2 files changed, 15 insertions, 8 deletions

diff --git a/python/sandcrawler/html_ingest.py b/python/sandcrawler/html_ingest.py
index acd336e..284461e 100644
--- a/python/sandcrawler/html_ingest.py
+++ b/python/sandcrawler/html_ingest.py
@@ -6,7 +6,7 @@ import json
 import datetime
 import argparse
 import xml.etree.ElementTree as ET
-from typing import List, Optional, Any
+from typing import List, Optional, Any, Tuple
 
 import trafilatura
 import pydantic
@@ -75,7 +75,7 @@ class IngestWebResult(pydantic.BaseModel):
         }
 
 
-def fix_transfer_encoding(file_meta: dict, resource: ResourceResult) -> (dict, ResourceResult):
+def fix_transfer_encoding(file_meta: dict, resource: ResourceResult) -> Tuple[dict, ResourceResult]:
     if file_meta['mimetype'] == 'application/gzip' and resource.cdx and resource.cdx.mimetype != 'application/gzip':
         print("transfer encoding not stripped: {}".format(resource.cdx.mimetype), file=sys.stderr)
         inner_body = gzip.decompress(resource.body)
@@ -233,7 +233,7 @@ def run_single(url: str, timestamp: Optional[str] = None, quick_mode: bool = Fal
         )
 
     html_doc = HTMLParser(html_resource.body)
-    html_biblio = html_extract_biblio(html_doc)
+    html_biblio = html_extract_biblio(url, html_doc)
     html_fulltext = html_extract_fulltext_teixml(html_resource.body)
     html_scope = html_guess_scope(url, html_doc, html_biblio, html_fulltext.get('tei_xml'))
     if html_scope not in ('article-fulltext', 'unknown'):
diff --git a/python/sandcrawler/html_metadata.py b/python/sandcrawler/html_metadata.py
index d3ca1b7..41157e0 100644
--- a/python/sandcrawler/html_metadata.py
+++ b/python/sandcrawler/html_metadata.py
@@ -232,11 +232,7 @@ class BiblioMetadata(pydantic.BaseModel):
     xml_fulltext_url: Optional[str]
 
 
-def html_extract_biblio(doc: HTMLParser) -> Optional[BiblioMetadata]:
-    """
-    TODO:
-    - meta dc.identifier: parse DOI
-    """
+def html_extract_biblio(doc_url: str, doc: HTMLParser) -> Optional[BiblioMetadata]:
 
     meta: Any = dict()
     head = doc.css_first("head")
@@ -262,6 +258,12 @@ def html_extract_biblio(doc: HTMLParser) -> Optional[BiblioMetadata]:
                         meta[field].append(val.attrs['content'])
                 break
 
+    # non-<meta> lookups
+    if not meta.get('xml_fulltext_url'):
+        val = head.css_first("link[rel='alternate'][type='application/xml']")
+        if val and val.attrs['href']:
+            meta['xml_fulltext_url'] = val.attrs['href']
+
     # TODO: replace with clean_doi() et al
     if meta.get('doi') and meta.get('doi').startswith('doi:'):
         meta['doi'] = meta['doi'][4:]
@@ -290,6 +292,11 @@ def html_extract_biblio(doc: HTMLParser) -> Optional[BiblioMetadata]:
         if release_type:
             meta['release_type'] = release_type
 
+    # resolve relative URLs
+    for key in ('pdf_fulltext_url', 'html_fulltext_url', 'xml_fulltext_url'):
+        if meta.get(key):
+            meta[key] = urllib.parse.urljoin(doc_url, meta[key])
+
     return BiblioMetadata(**meta)
 
 def load_adblock_rules() -> braveblock.Adblocker: