Update README.md
Browse files
README.md
CHANGED
@@ -24,7 +24,7 @@ This dataset repo contains only the queries and relevance judgements. The corpus
|
|
24 |
|
25 |
|
26 |
# Dataset Structure
|
27 |
-
To download the files: The queries can be found under `ciral-{lang}/topics` and are in `.tsv` formats with each line in the form:
|
28 |
|
29 |
```
|
30 |
qid\tquery
|
@@ -34,7 +34,25 @@ To download the files: The queries can be found under `ciral-{lang}/topics` and
|
|
34 |
```
|
35 |
qid Q0 docid relevance
|
36 |
```
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
37 |
|
|
|
38 |
## Citation
|
39 |
|
40 |
...
|
|
|
24 |
|
25 |
|
26 |
# Dataset Structure
|
27 |
+
1. To download the files: The queries can be found under `ciral-{lang}/topics` and are in `.tsv` formats with each line in the form:
|
28 |
|
29 |
```
|
30 |
qid\tquery
|
|
|
34 |
```
|
35 |
qid Q0 docid relevance
|
36 |
```
|
37 |
+
2. To access the dataset via `datasets`:
|
38 |
+
|
39 |
+
```
|
40 |
+
ciral_dataset = load_datasets("ciral/ciral", lang="hausa") #or swahili, somali, yoruba
|
41 |
+
|
42 |
+
for data in ciral_dataset['train']: # or 'test'
|
43 |
+
query_id = data['query_id']
|
44 |
+
query = data['query']
|
45 |
+
qrels = data['judgements']
|
46 |
+
|
47 |
+
for qrel in qrels:
|
48 |
+
docid = qrel['docid']
|
49 |
+
judgement = qrel['judgement']
|
50 |
+
text = qrel['text']
|
51 |
+
```
|
52 |
+
|
53 |
+
|
54 |
|
55 |
+
|
56 |
## Citation
|
57 |
|
58 |
...
|