[doc.pubtator] validity + request tracking: #183

- separation into 2 models, new one exclusively for tracking requests
- new validation approach to prevent return type confusion

mark2cure/document/models.py

@@ -121,11 +121,8 @@ class Pubtator(models.Model):
     document = models.ForeignKey(Document)
 
     kind = models.CharField(max_length=200, blank=True)
-    session_id = models.CharField(max_length=200, blank=True)
     content = models.TextField(blank=True, null=True)
 
-    request_count = models.IntegerField(default=0)
-    validate_cache = models.BooleanField(default=False)
 
     # Updated is also trigged during doc.valid_pubtator
     # so it's associated with when last polled or
@@ -139,18 +136,22 @@ class Meta:
     def __unicode__(self):
         return 'pubtator'
 
-    def valid(self):
-        # (TODO) This may return 2 different "types" check on
-        # implications of this discrepancy
-        if self.validate_cache:
-            return True
-
-        if self.session_id != '':
+    def is_valid(self):
+        """
+            Returns a boolean for if the pubtator is a valid state
+        """
+        if self.content is None:
             return False
 
-        if self.content is None:
+        if self.get_instance():
+            return True
+        else:
             return False
 
+    def get_instance(self):
+        """
+            Returns the pubtator BioC instance if valid or None
+        """
         try:
             r = BioCReader(source=self.content)
             r.read()
@@ -159,19 +160,38 @@ def valid(self):
             # If one of them doesn't validate leave
             return False
 
+
     def count_annotations(self):
-        if self.valid():
+        """
+            Returns an Integer count of all types of annotations, accross all sections for a pubtator response of any type.
+            If none are found or the document is invalid, return 0
+        """
+        instance = self.get_instance()
+        if instance:
             count = 0
-            reader = BioCReader(source=self.content)
-            reader.read()
-            for doc in reader.collection.documents:
+            for doc in instance.collection.documents:
                 for passage in doc.passages:
                     count += len(passage.annotations)
             return count
 
         else:
             return 0
 
+class PubtatorRequest(models.Model):
+    """
+        Pending jobs that have been submitted to Pubtator and are
+        awaiting completion
+    """
+    pubtator = models.ForeignKey(Pubtator)
+
+    fulfilled = models.BooleanField(default=False)
+    session_id = models.CharField(max_length=200, blank=True)
+    # The Number of times we've checked on the session_id
+    request_count = models.IntegerField(default=0)
+
+    updated = models.DateTimeField(auto_now=True)
+    created = models.DateTimeField(auto_now_add=True)
+
 
 class Section(models.Model):
     SECTION_KIND_CHOICE = (

mark2cure/document/signals.py

@@ -27,7 +27,10 @@ def pubtator_post_save(sender, instance, created, **kwargs):
             raise e
 
         session_id = re.findall(r'\d{4}-\d{4}-\d{4}-\d{4}', response.url)[0]
-        pubtator.session_id = session_id
-        pubtator.save()
+
+        PubtatorRequest.objects.get_or_create(
+            pubtator=pubtator,
+            fulfilled=False,
+            session_id=session_id)
 
 

mark2cure/document/tasks.py

@@ -113,11 +113,12 @@ def check_pubtator_health(self):
     # For all Pubtator models with content ensure it validates and cleanup the session_id and content
     for pubtator in Pubtator.objects.filter(content__isnull=False).all():
         # (TODO) Do robust checks for Pubtator object valid status
-        p_valid = pubtator.valid()
+        instance = pubtator.get_instance()
 
-        if p_valid:
+        if instance:
             # Association with the correct document
-            pubtator.document = Document.objects.get(document_id=p_valid.collection.documents[0].id)
+            doc_id = instance.collection.documents[0].id
+            pubtator.document = Document.objects.get(document_id=doc_id)
 
             # Prevents subsequent API calls
             pubtator.session_id = ''