Adding title/abstract + PRMU categories

README.md

@@ -42,6 +42,9 @@ This version of the dataset was produced by [(Boudin et al., 2016)][boudin-2016]
   * `lvl-4`: we abridge the input text from level 3 preprocessed documents using an unsupervised summarization technique.
     We keep the title and abstract and select the most content bearing sentences from the remaining contents.
 
+Titles and abstracts, collected from the [SciCorefCorpus](https://github.com/melsk125/SciCorefCorpus), are also provided.
+Details about how they were extracted and cleaned up can be found in [(Chaimongkol et al., 2014)][chaimongkol-2014]. 
+
 Reference keyphrases are provided in stemmed form (because they were provided like this for the test split in the competition).
 They are also categorized under the PRMU (<u>P</u>resent-<u>R</u>eordered-<u>M</u>ixed-<u>U</u>nseen) scheme as proposed in [(Boudin and Gallina, 2021)][boudin-2021].
 Text pre-processing (tokenization) is carried out using `spacy` (`en_core_web_sm` model) with a special rule to avoid splitting words with hyphens (e.g. graph-based is kept as one token).
@@ -50,13 +53,15 @@ Details about the process can be found in `prmu.py`.
 
 ## Content and statistics
 
-The dataset is divided into the following three splits:
+The dataset is divided into the following two splits:
 
 | Split      | # documents | #words | # keyphrases | % Present | % Reordered | % Mixed | % Unseen |
 | :--------- |------------:|-------:|-------------:|----------:|------------:|--------:|---------:|
 | Train      |         144 |      - |            - |         - |           - |       - |        - |
 | Test       |         100 |      - |            - |         - |           - |       - |        - |
 
+Statistics (#words, PRMU distributions) are computed using the title/abstract and not the full text of scientific papers.
+
 The following data fields are available :
 
 - **id**: unique identifier of the document.
@@ -71,9 +76,12 @@ The following data fields are available :
 
 ## References
 
-- (Kim et al., 2010). Su Nam Kim, Olena Medelyan, Min-Yen Kan, and Timothy Baldwin. 2010.
+- (Kim et al., 2010) Su Nam Kim, Olena Medelyan, Min-Yen Kan, and Timothy Baldwin. 2010.
   [SemEval-2010 Task 5 : Automatic Keyphrase Extraction from Scientific Articles][kim-2010].
   In Proceedings of the 5th International Workshop on Semantic Evaluation, pages 21–26, Uppsala, Sweden. Association for Computational Linguistics.
+- (Chaimongkol et al., 2014) Panot Chaimongkol, Akiko Aizawa, and Yuka Tateisi. 2014.
+  [Corpus for Coreference Resolution on Scientific Papers][chaimongkol-2014].
+  In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC'14), pages 3187–3190, Reykjavik, Iceland. European Language Resources Association (ELRA).
 - (Boudin et al., 2016) Florian Boudin, Hugo Mougard, and Damien Cram. 2016.
   [How Document Pre-processing affects Keyphrase Extraction Performance][boudin-2016].
   In Proceedings of the 2nd Workshop on Noisy User-generated Text (WNUT), pages 121–128, Osaka, Japan. The COLING 2016 Organizing Committee.
@@ -82,5 +90,6 @@ The following data fields are available :
   In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 4185–4193, Online. Association for Computational Linguistics.
 
 [kim-2010]: https://aclanthology.org/S10-1004/
+[chaimongkol-2014]: https://aclanthology.org/L14-1259/
 [boudin-2016]: https://aclanthology.org/W16-3917/
 [boudin-2021]: https://aclanthology.org/2021.naacl-main.330/

prmu.py

@@ -0,0 +1,100 @@
+# -*- coding: utf-8 -*-
+
+import sys
+import json
+import spacy
+
+from nltk.stem.snowball import SnowballStemmer as Stemmer
+
+nlp = spacy.load("en_core_web_sm")
+
+# https://spacy.io/usage/linguistic-features#native-tokenizer-additions
+
+from spacy.lang.char_classes import ALPHA, ALPHA_LOWER, ALPHA_UPPER
+from spacy.lang.char_classes import CONCAT_QUOTES, LIST_ELLIPSES, LIST_ICONS
+from spacy.util import compile_infix_regex
+
+# Modify tokenizer infix patterns
+infixes = (
+    LIST_ELLIPSES
+    + LIST_ICONS
+    + [
+        r"(?<=[0-9])[+\-\*^](?=[0-9-])",
+        r"(?<=[{al}{q}])\.(?=[{au}{q}])".format(
+            al=ALPHA_LOWER, au=ALPHA_UPPER, q=CONCAT_QUOTES
+        ),
+        r"(?<=[{a}]),(?=[{a}])".format(a=ALPHA),
+        # ✅ Commented out regex that splits on hyphens between letters:
+        # r"(?<=[{a}])(?:{h})(?=[{a}])".format(a=ALPHA, h=HYPHENS),
+        r"(?<=[{a}0-9])[:<>=/](?=[{a}])".format(a=ALPHA),
+    ]
+)
+
+infix_re = compile_infix_regex(infixes)
+nlp.tokenizer.infix_finditer = infix_re.finditer
+
+
+def contains(subseq, inseq):
+    return any(inseq[pos:pos + len(subseq)] == subseq for pos in range(0, len(inseq) - len(subseq) + 1))
+
+
+def find_pmru(tok_title, tok_text, tok_kp):
+    """Find PRMU category of a given keyphrase."""
+
+    # if kp is present
+    if contains(tok_kp, tok_title) or contains(tok_kp, tok_text):
+        return "P"
+
+    # if kp is considered as absent
+    else:
+
+        # find present and absent words
+        present_words = [w for w in tok_kp if w in tok_title or w in tok_text]
+
+        # if "all" words are present
+        if len(present_words) == len(tok_kp):
+            return "R"
+        # if "some" words are present
+        elif len(present_words) > 0:
+            return "M"
+        # if "no" words are present
+        else:
+            return "U"
+
+
+if __name__ == '__main__':
+
+    data = []
+
+    # read the dataset
+    with open(sys.argv[1], 'r') as f:
+        # loop through the documents
+        for line in f:
+            doc = json.loads(line.strip())
+
+            print(doc['id'])
+
+            title_spacy = nlp(doc['title'])
+            abstract_spacy = nlp(doc['abstract'])
+
+            title_tokens = [token.text for token in title_spacy]
+            abstract_tokens = [token.text for token in abstract_spacy]
+
+            title_stems = [Stemmer('porter').stem(w.lower()) for w in title_tokens]
+            abstract_stems = [Stemmer('porter').stem(w.lower()) for w in abstract_tokens]
+
+            keyphrases_stems = []
+            for keyphrase in doc['keyphrases']:
+                keyphrases_stems.append(keyphrase.split())
+
+            prmu = [find_pmru(title_stems, abstract_stems, kp) for kp in keyphrases_stems]
+
+            if doc['prmu'] != prmu:
+                print("PRMU categories are not identical!")
+
+            doc['prmu'] = prmu
+            data.append(json.dumps(doc))
+
+    # write the json
+    with open(sys.argv[2], 'w') as o:
+        o.write("\n".join(data))

test.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:5fda3666bcdbd4d6dd8b586864487132933734e0ed6e6aff6c758928b41ac46d
-size 11110282
+oid sha256:b2d985cd322c7ed8019d9a88ff3cbd92652bac155cb94fe20883e9ef235150ac
+size 11239360

train.jsonl

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ff5128cfe97a99cbc52b70d905cf731930f33072e6ebc4f7740b56e80818c8e9
-size 15885424
+oid sha256:45f7e772b653b3e7764857665129af800d190cf2e8801ee12fc60731b40404e6
+size 16057856