Updated script

README.md

@@ -39,15 +39,15 @@ This dataset repo contains only the queries and relevance judgements. The corpus
    ```
    ciral_dataset = load_dataset("ciral/ciral", "hausa") #or swahili, somali, yoruba
 
-   for data in ciral_dataset['train']:  # or 'test'
-     query_id = data['query_id']
-     query = data['query']
-     qrels = data['judgements']
-
-     for qrel in qrels:
-       docid = qrel['docid']
-       judgement = qrel['judgement']
-       text = qrel['text']
+   for data in ciral_data['train']:  # or 'test'
+      query_id = data['query_id']
+      query = data['query']
+      pos_qrels = data['positive_passages']
+      neg_qrels = data['negative_passages']
+
+      for qrel in pos_qrels:
+         docid = qrel['docid']
+         text = qrel['text']
    ```
 
 

ciral-yoruba/qrels/qrels.ciral-yoruba-train.tsv

@@ -57,8 +57,6 @@
 5	Q0	BBC#1268#1	0
 5	Q0	BBC#1268#2	0
 5	Q0	ALAROYE#961#2	0
-11	Q0	BBC#1381#19	1
-11	Q0	BBC#1381#20	1
 11	Q0	BBC#4159#2	1
 11	Q0	BBC#724#3	1
 11	Q0	BBC#724#4	1

ciral.py

@@ -90,9 +90,16 @@ class CIRAL(datasets.GeneratorBasedBuilder):
             {
                 "query_id": datasets.Value("string"),
                 "query": datasets.Value("string"),
-                "judgements": [{
+                # "judgements": [{
+                #     "docid": datasets.Value("string"),
+                #     "judgement": datasets.Value("string"),
+                #     "text": datasets.Value("string")}]
+                "positive_passages": [{
+                    'docid': datasets.Value("string"),
+                    'text': datasets.Value("string"),
+                }],
+                "negative_passages": [{
                     "docid": datasets.Value("string"),
-                    "judgement": datasets.Value("string"),
                     "text": datasets.Value("string")}]
             }
         )
@@ -137,13 +144,25 @@ class CIRAL(datasets.GeneratorBasedBuilder):
         queries = load_queries(query_file)
         qrels = load_qrels(qrel_file)
         for query_id in queries:
+
+            positive_docids = [docid for docid, judgement in qrels[query_id].items() if judgement==1]
+            negative_docids = [docid for docid, judgement in qrels[query_id].items() if judgement==0]
+
             data = {}
             data['query_id'] = query_id
             data['query'] = queries[query_id]
-            data['judgements'] = [{
+            data['positive_passages'] = [{
+                'docid': docid,
+                'text': docid2doc[docid]
+            } for docid in positive_docids if docid in docid2doc]
+            data['negative_passages'] = [{
                 'docid': docid,
-                'judgement': judgement,
                 'text': docid2doc[docid]
-            } for docid, judgement in qrels[query_id].items()]
+            } for docid in negative_docids if docid in docid2doc]
+            # data['judgements'] = [{
+            #     'docid': docid,
+            #     'judgement': judgement,
+            #     'text': docid2doc[docid]
+            # } for docid, judgement in qrels[query_id].items()]
 
             yield query_id, data