update ingest request schema

This is mostly changing ingest_type from 'file' to 'pdf', and adding
'link_source'/'link_source_id', plus some small cleanups.

3 files changed, 30 insertions, 8 deletions

diff --git a/python/fatcat_tools/importers/ingest.py b/python/fatcat_tools/importers/ingest.py
index a4258a38..bd5713c3 100644
--- a/python/fatcat_tools/importers/ingest.py
+++ b/python/fatcat_tools/importers/ingest.py
@@ -132,10 +132,15 @@ class IngestFileResultImporter(EntityImporter):
         )
         if fatcat and fatcat.get('edit_extra'):
             fe.edit_extra = fatcat['edit_extra']
+        else:
+            fe.edit_extra = dict()
         if request.get('ingest_request_source'):
-            if not fe.edit_extra:
-                fe.edit_extra = dict()
             fe.edit_extra['ingest_request_source'] = request['ingest_request_source']
+        if request.get('link_source') and request.get('link_source_id'):
+            fe.edit_extra['link_source'] = request['link_source']
+            fe.edit_extra['link_source_id'] = request['link_source_id']
+        if not fe.edit_extra:
+            fe.edit_extra = None
         return fe
 
     def try_update(self, fe):
diff --git a/python/fatcat_tools/transforms/ingest.py b/python/fatcat_tools/transforms/ingest.py
index 988f80a2..e08d56b8 100644
--- a/python/fatcat_tools/transforms/ingest.py
+++ b/python/fatcat_tools/transforms/ingest.py
@@ -1,7 +1,7 @@
 
 from .elasticsearch import release_to_elasticsearch
 
-def release_ingest_request(release, oa_only=False, ingest_request_source='fatcat', ingest_type='pdf'):
+def release_ingest_request(release, oa_only=False, ingest_request_source='fatcat', ingest_type=None):
     """
     Takes a full release entity object and returns an ingest request (as dict),
     or None if it seems like this release shouldn't be ingested.
@@ -12,7 +12,7 @@ def release_ingest_request(release, oa_only=False, ingest_request_source='fatcat
     The 'oa_only' boolean flag indicates that we should only return an ingest
     request if we have reason to believe this is an OA release (or, eg, in
     arxiv or pubmed central). Respecting this flag means we are likely to miss
-    a lot of "hybrid" and "bronze" content, but could reduce load
+    a lot of "hybrid" and "bronze" content, but could reduce crawl load
     significantly.
 
     The type of the ingest request may depend on release type and container
@@ -25,14 +25,22 @@ def release_ingest_request(release, oa_only=False, ingest_request_source='fatcat
 
     # generate a URL where we expect to find fulltext
     url = None
+    link_source = None
+    link_source_id = None
     if release.ext_ids.arxiv:
         url = "https://arxiv.org/pdf/{}.pdf".format(release.ext_ids.arxiv)
+        link_source = "arxiv"
+        link_source_id = release.ext_ids.arxiv
     elif release.ext_ids.doi:
         url = "https://doi.org/{}".format(release.ext_ids.doi)
-    elif release.ext_ids.pmcid:
+        link_source = "doi"
+        link_source_id = release.ext_ids.doi
+    elif release.ext_ids.pmcid and release.ext_ids.pmid:
         # TODO: how to tell if an author manuscript in PMC vs. published?
         #url = "https://www.ncbi.nlm.nih.gov/pmc/articles/{}/pdf/".format(release.ext_ids.pmcid)
         url = "http://europepmc.org/backend/ptpmcrender.fcgi?accid={}&blobtype=pdf".format(release.ext_ids.pmcid)
+        link_source = "pubmed"
+        link_source_id = release.ext_ids.pmid
 
     if not url:
         return None
@@ -40,21 +48,30 @@ def release_ingest_request(release, oa_only=False, ingest_request_source='fatcat
     ext_ids = release.ext_ids.to_dict()
     ext_ids = dict([(k, v) for (k, v) in ext_ids.items() if v])
 
-    if oa_only and not ext_ids.get('arxiv') and not ext_ids.get('pmcid'):
+    if oa_only and link_source not in ('arxiv', 'pubmed'):
         es = release_to_elasticsearch(release)
         if not es['is_oa']:
             return None
 
+    # TODO: infer ingest type based on release_type or container metadata?
+    if not ingest_type:
+        ingest_type = 'pdf'
+
     ingest_request = {
         'ingest_type': ingest_type,
         'ingest_request_source': ingest_request_source,
         'base_url': url,
+        'release_stage': release.release_stage,
         'fatcat': {
-            'release_stage': release.release_stage,
             'release_ident': release.ident,
             'work_ident': release.work_id,
         },
         'ext_ids': ext_ids,
     }
+
+    if link_source and link_source_id:
+        ingest_request['link_source'] = link_source
+        ingest_request['link_source_id'] = link_source_id
+
     return ingest_request
 
diff --git a/python/fatcat_tools/workers/changelog.py b/python/fatcat_tools/workers/changelog.py
index e1a72217..863ad40a 100644
--- a/python/fatcat_tools/workers/changelog.py
+++ b/python/fatcat_tools/workers/changelog.py
@@ -223,7 +223,7 @@ class EntityUpdatesWorker(FatcatWorker):
                 # filter to "new" active releases with no matched files
                 if release.ident in new_release_ids:
                     ir = release_ingest_request(release, ingest_request_source='fatcat-changelog', oa_only=self.ingest_oa_only)
-                    if ir and ir['ingest_type'] == 'file' and not release.files:
+                    if ir and not release.files:
                         producer.produce(
                             self.ingest_file_request_topic,
                             json.dumps(ir).encode('utf-8'),