update metadata to track search settings

lisc/collect/counts.py

@@ -86,8 +86,16 @@ def collect_counts(terms_a, inclusions_a=None, exclusions_a=None, labels_a=None,
     ...                                           terms_b=[['attention'], ['perception']])
     """
 
+    # Initialize meta data object
+    meta_data = MetaData()
+
+    # Collect settings for URLs, and add them to the metadata object
+    settings = {'db' : db, 'field' : field}
+    settings.update(eutils_kwargs)
+    meta_data.add_settings(settings)
+
     # Get e-utils URLS object. Set retmax as 0, since not using UIDs for counts
-    urls = EUtils(db=db, retmax='0', field=field, retmode='xml', **eutils_kwargs, api_key=api_key)
+    urls = EUtils(**settings, retmax='0', retmode='xml', api_key=api_key)
 
     # Define the settings for the search utility, adding a default for datetype if not provided
     search_settings = ['db', 'retmax', 'retmode', 'field']
@@ -98,9 +106,6 @@ def collect_counts(terms_a, inclusions_a=None, exclusions_a=None, labels_a=None,
     urls.build_url('info', settings=['db'])
     urls.build_url('search', settings=search_settings + list(eutils_kwargs.keys()))
 
-    # Initialize meta data object
-    meta_data = MetaData()
-
     # Check for a Requester object to be passed in as logging, otherwise initialize
     req = logging if isinstance(logging, Requester) else \
         Requester(wait_time=get_wait_time(urls.authenticated),

lisc/collect/words.py

@@ -80,9 +80,17 @@ def collect_words(terms, inclusions=None, exclusions=None, labels=None,
         msg = 'Only the `pubmed` database is currently supported for words collection.'
         raise NotImplementedError(msg)
 
+    # Initialize meta data object
+    meta_data = MetaData()
+
+    # Collect settings for URLs, and add them to the metadata object
+    settings = {'db' : db, 'retmax' : retmax, 'field' : field,
+                'usehistory' : 'y' if usehistory else 'n'}
+    settings.update(eutils_kwargs)
+    meta_data.add_settings(settings)
+
     # Get EUtils URLS object, with desired settings, and build required utility URLs
-    urls = EUtils(db=db, retmax=retmax, usehistory='y' if usehistory else 'n',
-                  field=field, retmode='xml', **eutils_kwargs, api_key=api_key)
+    urls = EUtils(**settings, retmode='xml', api_key=api_key)
 
     # Define the settings for the search utility, adding a default for datetype if not provided
     search_settings = ['db', 'usehistory', 'retmax', 'retmode', 'field']
@@ -94,10 +102,6 @@ def collect_words(terms, inclusions=None, exclusions=None, labels=None,
     urls.build_url('search', settings=search_settings + list(eutils_kwargs.keys()))
     urls.build_url('fetch', settings=['db', 'retmode'])
 
-    # Initialize results & meta data
-    results = []
-    meta_data = MetaData()
-
     # Check for a Requester object to be passed in as logging, otherwise initialize
     req = logging if isinstance(logging, Requester) else \
         Requester(wait_time=get_wait_time(urls.authenticated),
@@ -111,7 +115,8 @@ def collect_words(terms, inclusions=None, exclusions=None, labels=None,
     inclusions = inclusions if inclusions else [[]] * len(terms)
     exclusions = exclusions if exclusions else [[]] * len(terms)
 
-    # Loop through all the terms
+    # Loop through all the terms, launch collection, and collect results
+    results = []
     for label, search, incl, excl in zip(labels, terms, inclusions, exclusions):
 
         # Collect term information and make search term argument

lisc/data/meta_data.py

@@ -17,6 +17,8 @@ class MetaData():
         Details of the requester object used for the data collection.
     db_info : dict
         Details of the database from which the data was accessed.
+    settings : dict
+        Details of any search settings that were used during the collection.
     log : list or None
         A log of requested URLs, if requests were logged.
     """
@@ -27,6 +29,7 @@ def __init__(self):
         self.date = None
         self.requester = None
         self.db_info = None
+        self.settings = None
         self.log = None
 
         self.get_date()
@@ -47,7 +50,7 @@ def as_dict(self):
         mt_dict = deepcopy(self.__dict__)
 
         # Unpack dictionary attributes to flatten dictionary
-        for label in ['requester', 'db_info']:
+        for label in ['requester', 'db_info', 'settings']:
             attr = mt_dict.pop(label)
             if attr:
                 for key, val in attr.items():
@@ -92,3 +95,14 @@ def add_db_info(self, db_info):
         """
 
         self.db_info = db_info
+
+    def add_settings(self, settings):
+        """Add search settings information to the MetaData object.
+
+        Parameters
+        ----------
+        settings : dict
+            Information about settings that were used during the data collection.
+        """
+
+        self.settings = settings