added datagroup function

fixed output cleanup which removes methodology, system or dataset if no data present. Was causing an error if the output was called more than once. This seems to interfere with the jcamp and rruff tests as those seem to assume those sections are present
chalklab · Oct 29, 2021 · c4fed82 · c4fed82
1 parent 2cc63b1
commit c4fed82
Show file tree

Hide file tree

Showing 2 changed files with 56 additions and 21 deletions.
diff --git a/example.py b/example.py
@@ -279,6 +279,10 @@
                  "dataarray": seriesy}]}
 example.dataseries([dataser1])
 
+datagrp1 = {"@id": "datagroup",
+            "label": "datagroup 1",
+            "ids": ["datapoint/1/", "datapoint/2/"]}
+example.datagroup([datagrp1])
 
 # add source
 src = {'citation': 'Chalk Research Group',

diff --git a/scidatalib/scidata.py b/scidatalib/scidata.py
@@ -577,6 +577,39 @@ def dataseries(self, series: list) -> list:
         self.meta['@graph']['scidata'] = scidata
         return new_series
 
+    def datagroup(self, group: list) -> list:
+        """Add one or more datapoints"""
+        new_group = []
+        scidata: dict = self.meta['@graph']['scidata']
+        dataset: dict = scidata['dataset']
+        if 'datagroup' in dataset.keys():
+            curr_group: list = dataset['datagroup']
+        else:
+            curr_group = []
+        for listentry in group:
+            item = self.__iterate_function(listentry)
+            item_noid = {k: item[k] for k in set(list(item.keys())) - {'@id'}}
+            matched_group = 0
+            for groupitem in curr_group:
+                group_item_noid = {
+                    k: groupitem[k] for k in set(
+                        list(
+                            groupitem.keys())) -
+                    {'@id'}}
+                if group_item_noid == item_noid:
+                    matched_group = groupitem
+            if matched_group:
+                new_group.append(matched_group)
+                self.uidindex.remove(item['@id'])
+            else:
+                new_group.append(item)
+                curr_group.append(item)
+
+        dataset['datagroup'] = curr_group
+        scidata['dataset'] = dataset
+        self.meta['@graph']['scidata'] = scidata
+        return new_group
+
     def sources(self, sources: list, replace=False) -> dict:
         """
         Add to or replace the source reference list
@@ -758,28 +791,26 @@ def output(self) -> dict:
         for key in list(self.meta['@graph']['scidata']):
             if not self.meta['@graph']['scidata'][key]:
                 del self.meta['@graph']['scidata'][key]
-        sects = ['methodology', 'system', 'dataset']
-        for sect in sects:
-            for key in list(self.meta['@graph']['scidata'][sect]):
-                if not self.meta['@graph']['scidata'][sect][key]:
-                    del self.meta['@graph']['scidata'][sect][key]
-        if not self.meta['@graph']['scidata'].get(
-                'methodology',
-                {}).get(
-                'aspects',
-                False):
-            del self.meta['@graph']['scidata']['methodology']
-        if not self.meta['@graph']['scidata'].get(
-                'system',
-                {}).get(
-                'facets',
-                False):
-            del self.meta['@graph']['scidata']['system']
-        if not self.meta['@graph']['scidata'].get(
-                'dataset', {}).get('datapoint', False):
+        if self.meta['@graph']['scidata'].get('methodology'):
+            if not self.meta['@graph']['scidata'].get(
+                    'methodology',
+                    {}).get(
+                    'aspects',
+                    False):
+                del self.meta['@graph']['scidata']['methodology']
+        if self.meta['@graph']['scidata'].get('system'):
+            if not self.meta['@graph']['scidata'].get(
+                    'system',
+                    {}).get(
+                    'facets',
+                    False):
+                del self.meta['@graph']['scidata']['system']
+        if self.meta['@graph']['scidata'].get('dataset'):
             if not self.meta['@graph']['scidata'].get(
-                    'dataset', {}).get('dataseries', False):
-                del self.meta['@graph']['scidata']['dataset']
+                    'dataset', {}).get('datapoint', False):
+                if not self.meta['@graph']['scidata'].get(
+                        'dataset', {}).get('dataseries', False):
+                    del self.meta['@graph']['scidata']['dataset']
         self.__addtoc()
 
         return self.meta