Updated dataset loading to include anotations

README.md

@@ -1,65 +1,66 @@
 ---
 language:
-- en
-- ar
-- es
-- fr
-- ru
-- hi
-- ms
-- sw
-- az
-- ko
-- pt
-- hy
-- th
-- uk
-- ur
-- sr
-- iw
-- ja
-- hr
-- tl
-- ky
-- vi
-- fa
-- tg
-- mg
-- nl
-- ne
-- uz
-- my
-- da
-- dz
-- id
-- is
-- tr
-- lo
-- sl
-- so
-- mn
-- bn
-- bs
-- ht
-- el
-- it
-- to
-- ka
-- sn
-- sq
-- zh
+  - en
+  - ar
+  - es
+  - fr
+  - ru
+  - hi
+  - ms
+  - sw
+  - az
+  - ko
+  - pt
+  - hy
+  - th
+  - uk
+  - ur
+  - sr
+  - iw
+  - ja
+  - hr
+  - tl
+  - ky
+  - vi
+  - fa
+  - tg
+  - mg
+  - nl
+  - ne
+  - uz
+  - my
+  - da
+  - dz
+  - id
+  - is
+  - tr
+  - lo
+  - sl
+  - so
+  - mn
+  - bn
+  - bs
+  - ht
+  - el
+  - it
+  - to
+  - ka
+  - sn
+  - sq
+  - zh
 pretty_name: BordIRlines
 multilinguality:
-- multilingual
+  - multilingual
 annotations_creators:
-- machine-generated
+  - human
+  - machine-generated
 language_creators:
-- found
+  - found
 source_datasets:
-- manestay/borderlines
+  - manestay/borderlines
 license: mit
 task_categories:
-- question-answering
+  - question-answering
 arxiv: 2410.01171
 ---
 
@@ -74,85 +75,110 @@ Each `doc` is a passage from a Wikipedia article.
 
 ### Languages
 
-The dataset includes docs and queries in the following __languages__:
-
-* `en`: English
-* `zht`: Traditional Chinese
-* `ar`: Arabic
-* `zhs`: Simplified Chinese
-* `es`: Spanish
-* `fr`: French
-* `ru`: Russian
-* `hi`: Hindi
-* `ms`: Malay
-* `sw`: Swahili
-* `az`: Azerbaijani
-* `ko`: Korean
-* `pt`: Portuguese
-* `hy`: Armenian
-* `th`: Thai
-* `uk`: Ukrainian
-* `ur`: Urdu
-* `sr`: Serbian
-* `iw`: Hebrew
-* `ja`: Japanese
-* `hr`: Croatian
-* `tl`: Tagalog
-* `ky`: Kyrgyz
-* `vi`: Vietnamese
-* `fa`: Persian
-* `tg`: Tajik
-* `mg`: Malagasy
-* `nl`: Dutch
-* `ne`: Nepali
-* `uz`: Uzbek
-* `my`: Burmese
-* `da`: Danish
-* `dz`: Dzongkha
-* `id`: Indonesian
-* `is`: Icelandic
-* `tr`: Turkish
-* `lo`: Lao
-* `sl`: Slovenian
-* `so`: Somali
-* `mn`: Mongolian
-* `bn`: Bengali
-* `bs`: Bosnian
-* `ht`: Haitian Creole
-* `el`: Greek
-* `it`: Italian
-* `to`: Tonga
-* `ka`: Georgian
-* `sn`: Shona
-* `sq`: Albanian
-* `control`: see below
+The dataset includes docs and queries in the following **languages**:
+
+- `en`: English
+- `zht`: Traditional Chinese
+- `ar`: Arabic
+- `zhs`: Simplified Chinese
+- `es`: Spanish
+- `fr`: French
+- `ru`: Russian
+- `hi`: Hindi
+- `ms`: Malay
+- `sw`: Swahili
+- `az`: Azerbaijani
+- `ko`: Korean
+- `pt`: Portuguese
+- `hy`: Armenian
+- `th`: Thai
+- `uk`: Ukrainian
+- `ur`: Urdu
+- `sr`: Serbian
+- `iw`: Hebrew
+- `ja`: Japanese
+- `hr`: Croatian
+- `tl`: Tagalog
+- `ky`: Kyrgyz
+- `vi`: Vietnamese
+- `fa`: Persian
+- `tg`: Tajik
+- `mg`: Malagasy
+- `nl`: Dutch
+- `ne`: Nepali
+- `uz`: Uzbek
+- `my`: Burmese
+- `da`: Danish
+- `dz`: Dzongkha
+- `id`: Indonesian
+- `is`: Icelandic
+- `tr`: Turkish
+- `lo`: Lao
+- `sl`: Slovenian
+- `so`: Somali
+- `mn`: Mongolian
+- `bn`: Bengali
+- `bs`: Bosnian
+- `ht`: Haitian Creole
+- `el`: Greek
+- `it`: Italian
+- `to`: Tonga
+- `ka`: Georgian
+- `sn`: Shona
+- `sq`: Albanian
+- `control`: see below
 
 The **control** language is English, and contains the queries for all 251 territories. In contrast, **en** is only the 38 territories which have an English-speaking claimant.
 
+### Annotations
+
+The dataset contains two types of relevance annotations:
+
+1. **Human Annotations**:
+
+   - Provided by three annotators for a subset of query-document pairs.
+   - Relevance is determined by majority vote across annotators.
+   - Territories are listed per annotator, capturing individual perspectives.
+
+2. **LLM Annotations**:
+   - Includes two modes:
+     - **Zero-shot**: Predictions without any task-specific examples.
+     - **Few-shot**: Predictions with a small number of task-specific examples.
+   - Default mode is **few-shot**.
+
 ## Systems
+
 We have processed retrieval results for these IR systems:
-* `openai`: OpenAI's model `text-embedding-3-large`, cosine similarity
-* `m3`: M3-embedding ([link](https://huggingface.co/BAAI/bge-m3)) (Chen et al., 2024)
+
+- `openai`: OpenAI's model `text-embedding-3-large`, cosine similarity
+- `m3`: M3-embedding ([link](https://huggingface.co/BAAI/bge-m3)) (Chen et al., 2024)
 
 ## Modes
+
 Considering a user query in language `l` on a territory `t`, there are 4 modes for the IR.
-* `qlang`: consider passages in `{l}`. This is monolingual IR (the default).
-* `qlang_en`: consider passages in either `{l, en}`.
-* `en`: consider passages in `{en}`.
-* `rel_langs`:  consider passages in all relevant languages to `t` + `en`, so `{l1, l2, ..., en}`. This is a set, so `en` will not be duplicated if it already is relevant.
+
+- `qlang`: consider passages in `{l}`. This is monolingual IR (the default).
+- `qlang_en`: consider passages in either `{l, en}`.
+- `en`: consider passages in `{en}`.
+- `rel_langs`: consider passages in all relevant languages to `t` + `en`, so `{l1, l2, ..., en}`. This is a set, so `en` will not be duplicated if it already is relevant.
 
 ## Dataset Structure
 
 ### Data Fields
 
 The dataset consists of the following fields:
-* `query_id (string)`: The id of the query.
-* `query (string)`: The query text as provided by the `queries.tsv` file.
-* `territory (string)`: The territory of the query hit.
-* `rank (int32)`: The rank of the document for the corresponding query.
-* `score (float32)`: The relevance score of the document as provided by the search engine or model.
-* `doc_id (string)`: The unique identifier of the article.
-* `doc_text (string)`: The full text of the corresponding article or document.
+
+- `query_id (string)`: The id of the query.
+- `query (string)`: The query text as provided by the `queries.tsv` file.
+- `territory (string)`: The territory of the query hit.
+- `rank (int32)`: The rank of the document for the corresponding query.
+- `score (float32)`: The relevance score of the document as provided by the search engine or model.
+- `doc_id (string)`: The unique identifier of the article.
+- `doc_text (string)`: The full text of the corresponding article or document.
+- `relevant_human (bool)`: Majority relevance determined by human annotators.
+- `territory_human (list[string])`: Territories as judged by human annotators.
+- `relevant_llm_zeroshot (bool)`: LLM zero-shot relevance prediction.
+- `relevant_llm_fewshot (bool)`: LLM few-shot relevance prediction.
 
 ### Download Structure
 
@@ -167,13 +193,17 @@ data/
 ...
   all_docs.json
   queries.tsv
+  human_annotations.tsv
+  llm_annotations.tsv
 ```
 
-* `queries.tsv`: Contains the list of query IDs and their associated query texts.
-* `all_docs.json`: JSON dict containing all docs. It is organized as a nested dict, with keys `lang`, and values another dict with keys `doc_id`, and values `doc_text`.
-* `{lang}\_query_hits.tsv`: A TSV file with relevance scores and hit ranks for queries.
+- `queries.tsv`: Contains the list of query IDs and their associated query texts.
+- `all_docs.json`: JSON dict containing all docs. It is organized as a nested dict, with keys `lang`, and values another dict with keys `doc_id`, and values `doc_text`.
+- `{lang}\_query_hits.tsv`: A TSV file with relevance scores and hit ranks for queries.
+- `human_annotations.tsv`: A TSV file with human relevance annotations.
+- `llm_annotations.tsv`: A TSV file with LLM relevance predictions.
 
-Currently, there are 50 langs * 1 system * 4 modes = 200 query hit TSV files.
+Currently, there are 50 langs _ 1 system _ 4 modes = 200 query hit TSV files.
 
 ## Example Usage
 
@@ -195,9 +225,16 @@ ds_oa_zhs2 = load_dataset("borderlines/bordirlines", "zhs", split="openai.qlang"
 ds_m3_zhs1 = load_dataset("borderlines/bordirlines", "zhs", split="m3.en")
 # load Dataset for Simplified Chinese, qlang mode, m3 embedding
 ds_m3_zhs2 = load_dataset("borderlines/bordirlines", "zhs", split="m3.qlang")
+
+# Load Dataset for English, relevant-only with human annotations
+ds_human_en = load_dataset("borderlines/bordirlines", "en", relevant_only=True, annotation_type="human")
+
+# Load Dataset for Simplified Chinese, few-shot LLM mode
+ds_llm_fewshot_zhs = load_dataset("borderlines/bordirlines", "zhs", relevant_only=True, annotation_type="llm", llm_mode="fewshot")
 ```
 
 ## Citation
+
 ```
 @misc{li2024bordirlines,
       title={BordIRlines: A Dataset for Evaluating Cross-lingual Retrieval-Augmented Generation},

bordirlines.py

@@ -95,6 +95,12 @@ class BordIRLinesDataset(datasets.GeneratorBasedBuilder):
         for lang in SUPPORTED_LANGUAGES
     ]
 
+    def __init__(self, *args, relevant_only=False, annotation_type=None, llm_mode="fewshot", **kwargs):
+        super().__init__(*args, **kwargs)
+        self.relevant_only = relevant_only
+        self.annotation_type = annotation_type
+        self.llm_mode = llm_mode  # Choose between "zeroshot" and "fewshot". Default: "fewshot".
+
     def _info(self):
         return datasets.DatasetInfo(
             description="IR Dataset for BordIRLines paper.",
@@ -109,6 +115,10 @@ class BordIRLinesDataset(datasets.GeneratorBasedBuilder):
                     "doc_id": datasets.Value("string"),
                     "doc_text": datasets.Value("string"),
                     "doc_lang": datasets.Value("string"),
+                    "relevant_human": datasets.Value("bool"),
+                    "territory_human": datasets.Sequence(datasets.Value("string")),
+                    "relevant_llm_zeroshot": datasets.Value("bool"),
+                    "relevant_llm_fewshot": datasets.Value("bool"),
                 }
             ),
         )
@@ -117,6 +127,8 @@ class BordIRLinesDataset(datasets.GeneratorBasedBuilder):
         base_url = self.config.data_root_dir
         queries_path = f"{base_url}/queries.tsv"
         docs_path = dl_manager.download_and_extract(f"{base_url}/all_docs.json")
+        human_annotations_path = dl_manager.download_and_extract(f"{base_url}/human_annotations.tsv")
+        llm_annotations_path = dl_manager.download_and_extract(f"{base_url}/llm_annotations.tsv")
 
         lang = self.config.language
 
@@ -131,6 +143,8 @@ class BordIRLinesDataset(datasets.GeneratorBasedBuilder):
                         "hits": f"{base_url}/{lang}/{system}/{mode}/{lang}_query_hits.tsv",
                         "docs": docs_path,
                         "queries": queries_path,
+                        "human_annotations": human_annotations_path,
+                        "llm_annotations": llm_annotations_path,
                     }
                 )
 
@@ -140,13 +154,15 @@ class BordIRLinesDataset(datasets.GeneratorBasedBuilder):
                         "hits_path": downloaded_data[source]["hits"],
                         "docs_path": downloaded_data[source]["docs"],
                         "queries_path": downloaded_data[source]["queries"],
+                        "human_annotations_path": downloaded_data[source]["human_annotations"],
+                        "llm_annotations_path": downloaded_data[source]["llm_annotations"],
                     },
                 )
                 splits.append(split)
 
         return splits
 
-    def _generate_examples(self, hits_path, docs_path, queries_path):
+    def _generate_examples(self, hits_path, docs_path, queries_path, human_annotations_path, llm_annotations_path):
         n_hits = self.config.n_hits
         queries_df = pd.read_csv(queries_path, sep="\t")
         query_map = dict(zip(queries_df["query_id"], queries_df["query_text"]))
@@ -156,6 +172,9 @@ class BordIRLinesDataset(datasets.GeneratorBasedBuilder):
         docs = load_json(docs_path)
 
         hits = pd.read_csv(hits_path, sep="\t")
+        human_annotations = pd.read_csv(human_annotations_path, sep="\t")
+        llm_annotations = pd.read_csv(llm_annotations_path, sep="\t")
+        
         if n_hits:
             hits = hits.groupby("query_id").head(n_hits)
 
@@ -164,12 +183,57 @@ class BordIRLinesDataset(datasets.GeneratorBasedBuilder):
         hits = hits.sort_values(by=["query_id_int", "rank"])
         hits = hits.drop(columns=["query_id_int"])
 
+        human_map = human_annotations.set_index(["query_id", "doc_id"]).to_dict(orient="index")
+        llm_map = llm_annotations.set_index(["query_id", "doc_id"]).to_dict(orient="index")
+
         for _, row in hits.iterrows():
             doc_id = row["doc_id"]
             doc_lang = row["doc_lang"]
             query_id = row["query_id"]
             query_text = query_map[query_id]
             query_lang = query_to_lang_map[query_id]
+
+            # Get Human Data
+            human_data = human_map.get((query_id, doc_id), {})
+            # Parse relevant_human_votes manually
+            raw_votes = human_data.get("relevant_human", "[]")
+            relevant_human_votes = [
+                True if v.strip() == "True" else False if v.strip() == "False" else False
+                for v in raw_votes.strip("[]").split(",")
+                if v.strip()
+            ]
+
+            # Parse territory_human manually
+            raw_territories = human_data.get("territory_human", "[]")
+            territory_human = [
+                v.strip().strip("'").strip('"')  # Remove extra quotes and whitespace
+                for v in raw_territories.strip("[]").split(",")
+                if v.strip()
+            ]
+
+            # Calculate majority relevance
+            majority_relevant_human = (
+                sum(relevant_human_votes) > len(relevant_human_votes) / 2 if relevant_human_votes else False
+            )
+
+
+            # Get LLM Data
+            llm_data = llm_map.get((query_id, doc_id), {})
+            relevant_llm = (
+                llm_data.get("relevant_fewshot", None)
+                if self.llm_mode == "fewshot"
+                else llm_data.get("relevant_zeroshot", None)
+            )
+            # Filtering logic
+            if self.relevant_only:
+                if self.annotation_type == "human" and not majority_relevant_human:
+                    continue
+                elif self.annotation_type == "llm" and not (relevant_llm is True):
+                    continue
+                elif not majority_relevant_human and not (relevant_llm is True):
+                    continue
+
+
             yield (
                 counter,
                 {
@@ -182,6 +246,10 @@ class BordIRLinesDataset(datasets.GeneratorBasedBuilder):
                     "doc_id": doc_id,
                     "doc_text": docs[doc_lang][doc_id],
                     "doc_lang": doc_lang,
+                    "relevant_human": majority_relevant_human,
+                    "territory_human": territory_human,
+                    "relevant_llm_zeroshot": llm_data.get("relevant_zeroshot", None),
+                    "relevant_llm_fewshot": llm_data.get("relevant_fewshot", None),
                 },
             )