fix readme

README.md

@@ -19,17 +19,24 @@ pretty_name: Analogy Question
 ### Dataset Summary
 This dataset contains 5 different word analogy questions used in [Analogy Language Model](https://aclanthology.org/2021.acl-long.280/).
 
-| name | Size (valid/test) | Num of choice | Num of relation group | Original Reference                                                         |
-|---------|------------------:|--------------:|----------------------:|:--------------------------------------------------------------------------:|
+- original analogy questions
+
+| name      | Size (valid/test) | Num of choice | Num of relation group | Original Reference                                                         |
+|-----------|------------------:|--------------:|----------------------:|:--------------------------------------------------------------------------:|
 | `sat_full`| -/374             | 5             | 2                     | [Turney (2005)](https://arxiv.org/pdf/cs/0508053.pdf)                      |
 | `sat`     | 37/337            | 5             | 2                     | [Turney (2005)](https://arxiv.org/pdf/cs/0508053.pdf)                      |
 | `u2`      | 24/228            | 5,4,3         | 9                     | [EnglishForEveryone](https://englishforeveryone.org/Topics/Analogies.html) |
 | `u4`      | 48/432            | 5,4,3         | 5                     | [EnglishForEveryone](https://englishforeveryone.org/Topics/Analogies.html) |
 | `google`  | 50/500            | 4             | 2                     | [Mikolov et al., (2013)](https://www.aclweb.org/anthology/N13-1090.pdf)    |
 | `bats`    | 199/1799          | 4             | 3                     | [Gladkova et al., (2016)](https://www.aclweb.org/anthology/N18-2017.pdf)   |
-| `semeval2012_relational_similarity`   | 78/-      | 3 | -             | [relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity)  |
-| `t_rex_relational_similarity`           | 467/467   | 6 | -             | [relbert/t_rex_relational_similarity](https://huggingface.co/datasets/relbert/t_rex_relational_similarity)  |
-| `conceptnet_relational_similarity`      | 546/570   | 6 | -             | [relbert/conceptnet_relational_similarity](https://huggingface.co/datasets/relbert/conceptnet_relational_similarity)  |
+
+- extra analogy questions
+
+| name                                | Size (valid/test)   | Num of choice (valid/test)   | Num of relation group (valid/test)   | Original Reference                                                                                                     |
+|:------------------------------------|:--------------------|:-----------------------------|:-------------------------------------|:-----------------------------------------------------------------------------------------------------------------------|
+| `semeval2012_relational_similarity` | 79/-                | 3/-                          | 79/-                                 | [relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity) |
+| `t_rex_relational_similarity`       | 496/183             | 74/48                        | 60/19                                | [relbert/t_rex_relational_similarity](https://huggingface.co/datasets/relbert/t_rex_relational_similarity)             |
+| `conceptnet_relational_similarity`  | 1112/1192           | 19/17                        | 18/16                                | [relbert/conceptnet_relational_similarity](https://huggingface.co/datasets/relbert/conceptnet_relational_similarity)   |
 
 ## Dataset Structure
 ### Data Instances

add_new_analogy.py

@@ -1,94 +1,113 @@
 import json
 import os
-from itertools import combinations, chain
+from itertools import combinations
 from random import shuffle, seed
+
+import pandas as pd
 from datasets import load_dataset
 
-# # create analogy from `relbert/semeval2012_relational_similarity`
-# data = load_dataset("relbert/semeval2012_relational_similarity", split="validation")
-# analogy_data = [{
-#     "stem": i['positives'][0], "choice": i["negatives"] + [i['positives'][1]], "answer": 2, "prefix": i["relation_type"]
-# } for i in data]
-# os.makedirs("dataset/semeval2012_relational_similarity", exist_ok=True)
-# with open("dataset/semeval2012_relational_similarity/valid.jsonl", "w") as f:
-#     f.write("\n".join([json.dumps(i) for i in analogy_data]))
 
+def get_stats(filename):
+    with open(filename) as f:
+        _data = [json.loads(i) for i in f.read().splitlines()]
+    return len(_data), list(set([len(i['choice']) for i in _data])), len(list(set([i['prefix'] for i in _data])))
+
+
+def lexical_overlap(word_a, word_b):
+    for a in word_a.split(" "):
+        for b in word_b.split(" "):
+            if a.lower() == b.lower():
+                return True
+    return False
+
+
+def create_analogy(_data, output_path, negative_per_relation, instance_per_relation=100):
+    # if os.path.exists(output_path):
+    #     return
+    df = _data.to_pandas()
+    analogy_data = []
+    for _, i in df.iterrows():
+        target = [(q.tolist(), c.tolist()) for q, c in combinations(i['positives'], 2)
+                  if not any(lexical_overlap(c[0], y) or lexical_overlap(c[1], y) for y in q)]
+        if len(target) == 0:
+            continue
+        if len(target) > instance_per_relation:
+            seed(42)
+            shuffle(target)
+            target = target[:instance_per_relation]
+        for m, (q, c) in enumerate(target):
+            negative = []
+            for r in df['relation_type']:
+                if r == i['relation_type']:
+                    continue
+                target_per_relation = [y.tolist() for y in df[df['relation_type'] == r]['positives'].values[0]]
+                shuffle(target_per_relation)
+                negative += target_per_relation[:negative_per_relation]
+            analogy_data.append({
+                "stem": q,
+                "choice": [c, c[::-1]] + negative,
+                "answer": 0,
+                "prefix": i["relation_type"]
+            })
+    os.makedirs(os.path.dirname(output_path), exist_ok=True)
+    with open(output_path, "w") as f:
+        f.write("\n".join([json.dumps(i) for i in analogy_data]))
+
+stat = []
+###################################################################
+# create analogy from `relbert/semeval2012_relational_similarity` #
+###################################################################
+if not os.path.exists("dataset/semeval2012_relational_similarity/valid.jsonl"):
+    data = load_dataset("relbert/semeval2012_relational_similarity", split="validation")
+    analogy_data = [{
+        "stem": i['positives'][0], "choice": i["negatives"] + [i['positives'][1]], "answer": 2, "prefix": i["relation_type"]
+    } for i in data]
+    os.makedirs("dataset/semeval2012_relational_similarity", exist_ok=True)
+    with open("dataset/semeval2012_relational_similarity/valid.jsonl", "w") as f:
+        f.write("\n".join([json.dumps(i) for i in analogy_data]))
+
+v_size, v_num_choice, v_relation_type = get_stats("dataset/semeval2012_relational_similarity/valid.jsonl")
+stat.append({
+    "name": "`semeval2012_relational_similarity`",
+    "Size (valid/test)": f"{v_size}/-",
+    "Num of choice (valid/test)": f"{','.join([str(n) for n in v_num_choice])}/-",
+    "Num of relation group (valid/test)": f"{v_relation_type}/-",
+    "Original Reference": "[relbert/semeval2012_relational_similarity](https://huggingface.co/datasets/relbert/semeval2012_relational_similarity)"
+})
+
+
+#############################################################
+# create analogy from `relbert/t_rex_relational_similarity` #
+#############################################################
+data = load_dataset("relbert/t_rex_relational_similarity", "filter_unified.min_entity_1_max_predicate_100", split="test")
+create_analogy(data, "dataset/t_rex_relational_similarity/test.jsonl", negative_per_relation=2)
+data = load_dataset("relbert/t_rex_relational_similarity", "filter_unified.min_entity_4_max_predicate_100", split="validation")
+create_analogy(data, "dataset/t_rex_relational_similarity/valid.jsonl", negative_per_relation=1)
+
+t_size, t_num_choice, t_relation_type = get_stats("dataset/t_rex_relational_similarity/test.jsonl")
+v_size, v_num_choice, v_relation_type = get_stats("dataset/t_rex_relational_similarity/valid.jsonl")
+stat.append({
+    "name": "`t_rex_relational_similarity`",
+    "Size (valid/test)": f"{v_size}/{t_size}",
+    "Num of choice (valid/test)": f"{','.join([str(n) for n in v_num_choice])}/{','.join([str(n) for n in t_num_choice])}",
+    "Num of relation group (valid/test)": f"{v_relation_type}/{t_relation_type}",
+    "Original Reference": "[relbert/t_rex_relational_similarity](https://huggingface.co/datasets/relbert/t_rex_relational_similarity)"
+})
 
-# # create analogy from `relbert/t_rex_relational_similarity`
-# data = load_dataset("relbert/t_rex_relational_similarity", "filter_unified.min_entity_1_max_predicate_100", split="test")
-# df = data.to_pandas()
-# df['negatives'] = [list(chain(
-#     *[[y.tolist() for y in x.tolist()] for x in df[df.relation_type != i]['positives'].tolist()] +
-#      [[y.tolist() for y in x.tolist()] for x in df[df.relation_type == i]['negatives'].tolist()])) for i in
-#     df['relation_type']]
-# analogy_data = []
-# for _, i in df.iterrows():
-#     if len(i['positives']) < 2:
-#         continue
-#     for m, (q, c) in enumerate(combinations(i['positives'], 2)):
-#         if m > 5:
-#             break
-#         negative = i['negatives']
-#         for n in range(6):
-#             seed(n)
-#             shuffle(negative)
-#             analogy_data.append({
-#                 "stem": q.tolist(), "choice": [c.tolist()] + negative[:5], "answer": 0, "prefix": i["relation_type"]
-#             })
-# os.makedirs("dataset/t_rex_relational_similarity", exist_ok=True)
-# with open("dataset/t_rex_relational_similarity/test.jsonl", "w") as f:
-#     f.write("\n".join([json.dumps(i) for i in analogy_data]))
-#
-# data = load_dataset("relbert/t_rex_relational_similarity", "filter_unified.min_entity_4_max_predicate_100", split="validation")
-# df = data.to_pandas()
-# df['negatives'] = [list(chain(
-#     *[[y.tolist() for y in x.tolist()] for x in df[df.relation_type != i]['positives'].tolist()] +
-#      [[y.tolist() for y in x.tolist()] for x in df[df.relation_type == i]['negatives'].tolist()])) for i in
-#     df['relation_type']]
-# analogy_data = []
-# for _, i in df.iterrows():
-#     if len(i['positives']) < 5:
-#         continue
-#     for m, (q, c) in enumerate(combinations(i['positives'], 2)):
-#         if m > 5:
-#             break
-#         negative = i['negatives']
-#         for n in range(3):
-#             seed(n)
-#             shuffle(negative)
-#             analogy_data.append({
-#                 "stem": q.tolist(), "choice": [c.tolist()] + negative[:5], "answer": 0, "prefix": i["relation_type"]
-#             })
-# os.makedirs("dataset/t_rex_relational_similarity", exist_ok=True)
-# with open("dataset/t_rex_relational_similarity/valid.jsonl", "w") as f:
-#     f.write("\n".join([json.dumps(i) for i in analogy_data]))
-#
-# # create analogy from `relbert/conceptnet_relational_similarity`
-# for s in ['test', 'validation']:
-#     data = load_dataset("relbert/conceptnet_relational_similarity", split=s)
-#     df = data.to_pandas()
-#     df['negatives'] = [list(chain(
-#         *[[y.tolist() for y in x.tolist()] for x in df[df.relation_type != i]['positives'].tolist()] +
-#          [[y.tolist() for y in x.tolist()] for x in df[df.relation_type == i]['negatives'].tolist()])) for i in
-#         df['relation_type']]
-#
-#     analogy_data = []
-#
-#     for _, i in df.iterrows():
-#
-#         if len(i['positives']) < 2:
-#             continue
-#         for m, (q, c) in enumerate(combinations(i['positives'], 2)):
-#             if m > 5:
-#                 break
-#             negative = i['negatives']
-#             for n in range(6):
-#                 seed(n)
-#                 shuffle(negative)
-#                 analogy_data.append({
-#                     "stem": q.tolist(), "choice": [c.tolist()] + negative[:5], "answer": 0, "prefix": i["relation_type"]
-#                 })
-#     print(len(analogy_data))
-#     os.makedirs("dataset/conceptnet_relational_similarity", exist_ok=True)
-#     with open(f"dataset/conceptnet_relational_similarity/{s if s == 'test' else 'valid'}.jsonl", "w") as f:
-#         f.write("\n".join([json.dumps(i) for i in analogy_data]))
+##################################################################
+# create analogy from `relbert/conceptnet_relational_similarity` #
+##################################################################
+data = load_dataset("relbert/conceptnet_relational_similarity", split="test")
+create_analogy(data, "dataset/conceptnet_relational_similarity/test.jsonl", negative_per_relation=1)
+data = load_dataset("relbert/conceptnet_relational_similarity", split="validation")
+create_analogy(data, "dataset/conceptnet_relational_similarity/valid.jsonl", negative_per_relation=1)
+t_size, t_num_choice, t_relation_type = get_stats("dataset/conceptnet_relational_similarity/test.jsonl")
+v_size, v_num_choice, v_relation_type = get_stats("dataset/conceptnet_relational_similarity/valid.jsonl")
+stat.append({
+    "name": "`conceptnet_relational_similarity`",
+    "Size (valid/test)": f"{v_size}/{t_size}",
+    "Num of choice (valid/test)": f"{','.join([str(n) for n in v_num_choice])}/{','.join([str(n) for n in t_num_choice])}",
+    "Num of relation group (valid/test)": f"{v_relation_type}/{t_relation_type}",
+    "Original Reference": "[relbert/conceptnet_relational_similarity](https://huggingface.co/datasets/relbert/conceptnet_relational_similarity)"
+})
+print(pd.DataFrame(stat).to_markdown(index=False))