Merge branch 'main' of hf.co:datasets/philipphager/baidu-ultr_baidu-mlm-ctr

README.md

@@ -50,16 +50,32 @@ Each row of the click / annotation dataset contains the following attributes. Us
 |------------------------------|----------------|-------------|
 | query_id                     | string         | Baidu query_id |
 | query_md5                    | string         | MD5 hash of query text |
-| url_md5                      | List[string]   | MD5 hash of document url, most reliable document identifier |
+| query                        | List[int32]    | List of query tokens |
+| query_length                 | int32          | Number of query tokens |
+| n                            | int32          | Number of documents for current query, useful for padding |
+| url_md5                      | List[string]   | MD5 hash of document URL, most reliable document identifier |
 | text_md5                     | List[string]   | MD5 hash of document title and abstract |
-| query_document_embedding     | Tensor[float16]| BERT CLS token |
+| title                        | List[List[int32]]    | List of tokens for document titles |
+| abstract                     | List[List[int32]]    | List of tokens for document abstracts |
+| query_document_embedding     | Tensor[Tensor[float16]]| BERT CLS token |
 | click                        | Tensor[int32]  | Click / no click on a document |
-| n                            | int32          | Number of documents for current query, useful for padding |
 | position                     | Tensor[int32]  | Position in ranking (does not always match original item position) |
-| media_type                   | Tensor[int32]  | Document type (label encoding recommended as ids do not occupy a continous integer range) |
-| displayed_time               | Tensor[float32]| Seconds a document was displayed on screen |
-| serp_height                  | Tensor[int32]  | Pixel height of a document on screen |
-| slipoff_count_after_click    | Tensor[int32]  | Number of times a document was scrolled off screen after previously clicking on it |
+| media_type                   | Tensor[int32]  | Document type (label encoding recommended as IDs do not occupy a continuous integer range) |
+| displayed_time               | Tensor[float32]| Seconds a document was displayed on the screen |
+| serp_height                  | Tensor[int32]  | Pixel height of a document on the screen |
+| slipoff_count_after_click    | Tensor[int32]  | Number of times a document was scrolled off the screen after previously clicking on it |
+| bm25                         | Tensor[float32]  | BM25 score for documents |
+| title_bm25                   | Tensor[float32]  | BM25 score for document titles |
+| abstract_bm25                | Tensor[float32]  | BM25 score for document abstracts |
+| tf_idf                       | Tensor[float32]  | TF-IDF score for documents |
+| tf                           | Tensor[float32]  | Term frequency for documents |
+| idf                          | Tensor[float32]  | Inverse document frequency for documents |
+| ql_jelinek_mercer_short      | Tensor[float32]  | Query likelihood score for documents using Jelinek-Mercer smoothing (alpha = 0.1) |
+| ql_jelinek_mercer_long       | Tensor[float32]  | Query likelihood score for documents using Jelinek-Mercer smoothing (alpha = 0.7) |
+| ql_dirichlet                 | Tensor[float32]  | Query likelihood score for documents using Dirichlet smoothing (lambda = 128) |
+| document_length              | Tensor[int32]    | Length of documents |
+| title_length                 | Tensor[int32]    | Length of document titles |
+| abstract_length              | Tensor[int32]    | Length of document abstracts |
 
 
 ### Expert annotation dataset
@@ -67,11 +83,28 @@ Each row of the click / annotation dataset contains the following attributes. Us
 |------------------------------|----------------|-------------|
 | query_id                     | string         | Baidu query_id |
 | query_md5                    | string         | MD5 hash of query text |
-| text_md5                     | List[string]   | MD5 hash of document title and abstract |
-| query_document_embedding     | Tensor[float16]| BERT CLS token |
-| label                        | Tensor[int32]  | Relevance judgment on a scale from 0 (bad) to 4 (excellent) |
-| n                            | int32          | Number of documents for current query, useful for padding |
+| query                        | List[int32]    | List of query tokens |
+| query_length                 | int32          | Number of query tokens |
 | frequency_bucket             | int32          | Monthly frequency of query (bucket) from 0 (high frequency) to 9 (low frequency) |
+| n                            | int32          | Number of documents for current query, useful for padding |
+| url_md5                      | List[string]   | MD5 hash of document URL, most reliable document identifier |
+| text_md5                     | List[string]   | MD5 hash of document title and abstract |
+| title                        | List[List[int32]]    | List of tokens for document titles |
+| abstract                     | List[List[int32]]    | List of tokens for document abstracts |
+| query_document_embedding     | Tensor[Tensor[float16]]  | BERT CLS token |
+| label                        | Tensor[int32]    | Relevance judgments on a scale from 0 (bad) to 4 (excellent) |
+| bm25                         | Tensor[float32]  | BM25 score for documents |
+| title_bm25                   | Tensor[float32]  | BM25 score for document titles |
+| abstract_bm25                | Tensor[float32]  | BM25 score for document abstracts |
+| tf_idf                       | Tensor[float32]  | TF-IDF score for documents |
+| tf                           | Tensor[float32]  | Term frequency for documents |
+| idf                          | Tensor[float32]  | Inverse document frequency for documents |
+| ql_jelinek_mercer_short      | Tensor[float32]  | Query likelihood score for documents using Jelinek-Mercer smoothing (alpha = 0.1) |
+| ql_jelinek_mercer_long       | Tensor[float32]  | Query likelihood score for documents using Jelinek-Mercer smoothing (alpha = 0.7) |
+| ql_dirichlet                 | Tensor[float32]  | Query likelihood score for documents using Dirichlet smoothing (lambda = 128) |
+| document_length              | Tensor[int32]    | Length of documents |
+| title_length                 | Tensor[int32]    | Length of document titles |
+| abstract_length              | Tensor[int32]    | Length of document abstracts |
 
 ## Example PyTorch collate function
 Each sample in the dataset is a single query with multiple documents.

baidu-ultr_baidu-mlm-ctr.py

@@ -53,30 +53,69 @@ class BaiduUltrBuilder(datasets.GeneratorBasedBuilder):
 
     CLICK_FEATURES = Features(
         {
+            ### Query features
             "query_id": Value("string"),
             "query_md5": Value("string"),
+            "query": Sequence(Value("int32")),
+            "query_length": Value("int32"),
+            "n": Value("int32"),
+            ### Doc features
             "url_md5": Sequence(Value("string")),
             "text_md5": Sequence(Value("string")),
+            "title": Sequence(Sequence(Value("int32"))),
+            "abstract": Sequence(Sequence(Value("int32"))),
             "query_document_embedding": Array2D((None, 768), "float16"),
             "click": Sequence(Value("int32")),
-            "n": Value("int32"),
+            ### SERP features
             "position": Sequence(Value("int32")),
             "media_type": Sequence(Value("int32")),
             "displayed_time": Sequence(Value("float32")),
             "serp_height": Sequence(Value("int32")),
             "slipoff_count_after_click": Sequence(Value("int32")),
+            ### LTR features
+            "bm25": Sequence(Value("float32")),
+            "title_bm25": Sequence(Value("float32")),
+            "abstract_bm25": Sequence(Value("float32")),
+            "tf_idf": Sequence(Value("float32")),
+            "tf": Sequence(Value("float32")),
+            "idf": Sequence(Value("float32")),
+            "ql_jelinek_mercer_short": Sequence(Value("float32")),
+            "ql_jelinek_mercer_long": Sequence(Value("float32")),
+            "ql_dirichlet": Sequence(Value("float32")),
+            "document_length": Sequence(Value("int32")),
+            "title_length": Sequence(Value("int32")),
+            "abstract_length": Sequence(Value("int32")),
         }
     )
 
     ANNOTATION_FEATURES = Features(
         {
+            ### Query features
             "query_id": Value("string"),
             "query_md5": Value("string"),
-            "text_md5": Value("string"),
+            "query": Sequence(Value("int32")),
+            "query_length": Value("int32"),
+            "frequency_bucket": Value("int32"),
+            "n": Value("int32"),
+            ### Doc features
+            "text_md5": Sequence(Value("string")),
+            "title": Sequence(Sequence(Value("int32"))),
+            "abstract": Sequence(Sequence(Value("int32"))),
             "query_document_embedding": Array2D((None, 768), "float16"),
             "label": Sequence(Value("int32")),
-            "n": Value("int32"),
-            "frequency_bucket": Value("int32"),
+            ### LTR features
+            "bm25": Sequence(Value("float32")),
+            "title_bm25": Sequence(Value("float32")),
+            "abstract_bm25": Sequence(Value("float32")),
+            "tf_idf": Sequence(Value("float32")),
+            "tf": Sequence(Value("float32")),
+            "idf": Sequence(Value("float32")),
+            "ql_jelinek_mercer_short": Sequence(Value("float32")),
+            "ql_jelinek_mercer_long": Sequence(Value("float32")),
+            "ql_dirichlet": Sequence(Value("float32")),
+            "document_length": Sequence(Value("int32")),
+            "title_length": Sequence(Value("int32")),
+            "abstract_length": Sequence(Value("int32")),
         }
     )
 
@@ -108,19 +147,34 @@ class BaiduUltrBuilder(datasets.GeneratorBasedBuilder):
             query_columns = [
                 "query_id",
                 "query_md5",
+                "query",
+                "query_length",
             ]
 
             agg_columns = [
-                "query_md5",
                 "url_md5",
                 "text_md5",
-                "position",
-                "click",
+                "title",
+                "abstract",
                 "query_document_embedding",
+                "click",
+                "position",
                 "media_type",
                 "displayed_time",
                 "serp_height",
                 "slipoff_count_after_click",
+                "bm25",
+                "title_bm25",
+                "abstract_bm25",
+                "tf_idf",
+                "tf",
+                "idf",
+                "ql_jelinek_mercer_short",
+                "ql_jelinek_mercer_long",
+                "ql_dirichlet",
+                "document_length",
+                "title_length",
+                "abstract_length",
             ]
 
             return [
@@ -146,12 +200,28 @@ class BaiduUltrBuilder(datasets.GeneratorBasedBuilder):
             query_columns = [
                 "query_id",
                 "query_md5",
+                "query",
+                "query_length",
                 "frequency_bucket",
             ]
             agg_columns = [
                 "text_md5",
-                "label",
+                "title",
+                "abstract",
                 "query_document_embedding",
+                "label",
+                "bm25",
+                "title_bm25",
+                "abstract_bm25",
+                "tf_idf",
+                "tf",
+                "idf",
+                "ql_jelinek_mercer_short",
+                "ql_jelinek_mercer_long",
+                "ql_dirichlet",
+                "document_length",
+                "title_length",
+                "abstract_length",
             ]
 
             return [