Simplify paper IDs, include raw proofs and paper tags

prooflang.py

@@ -11,7 +11,6 @@
 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
 # See the License for the specific language governing permissions and
 # limitations under the License.
-# TODO: Address all TODOs and remove all explanatory comments
 """The ProofLang Corpus of arXiv Proofs"""
 
 
@@ -21,18 +20,16 @@ import os
 import datasets
 
 
-# TODO: Add BibTeX citation
-# Find for instance the citation on arxiv or on the dataset repo/website
 _CITATION = """\
-@unpublished{prooflang:dataset,
-title = {ProofLang: the Language of arXiv Proofs},
-author = {Henry Hammer and Nanako Noda and Christopher A. Stone},
-year = {2023}
+@inproceedings{prooflang:dataset,
+title = "{ProofLang: the Language of arXiv Proofs}",
+booktitle = "{Intelligent Computer Mathematics (CICM 2023)}",
+author = "{Henry Hammer and Nanako Noda and Christopher A. Stone}",
+year = {2023},
+note = {To appear}
 }
 """
 
-# TODO: Add description of the dataset here
-# You can copy an official description
 _DESCRIPTION = """\ The ProofLang Corpus includes over three million
 English-language proofs—558 million words—mechanically extracted from the papers
 (Math, CS, Physics, etc.) posted on arXiv.org between 1992 and 2020. The focus
@@ -44,30 +41,21 @@ write informal proofs. It is also amenable to statistical analyses and to
 experiments with  Natural Language Processing (NLP) techniques.
 """
 
-# TODO: Add a link to an official homepage for the dataset here
-_HOMEPAGE = ""
+_HOMEPAGE = "https://huggingface.co/datasets/proofcheck/prooflang"
 
-# TODO: Add the licence for the dataset here if you can find it
 _LICENSE = "CC-BY 4.0"
 
-# TODO: Add link to the official dataset URLs here
-# The HuggingFace Datasets library doesn't host the datasets but only points to the original files.
-# This can be an arbitrary nested dict/list of URLs (see below in `_split_generators` method)
 _URLS = {
-    # "proofs": "https://huggingface.co/great-new-dataset-proofs.zip",
-    # "sentences": "https://huggingface.co/great-new-dataset-sentences.zip",
-    # "proofs": "cleanproofs19.tsv",
-    # "sentences": "sent19.tsv"
     "proofs": "proofs.zip",
     "sentences": "sentences.zip",
+    "raw": "raw.zip",
+    "tags": "tags.zip",
 }
 
-
-# TODO: Name of the dataset usually match the script name with CamelCase instead of snake_case
 class ArxivProofs(datasets.GeneratorBasedBuilder):
-    """TODO: Short description of my dataset."""
+    """English text from proofs found in arXiv preprints."""
 
-    VERSION = datasets.Version("0.5.2")
+    VERSION = datasets.Version("0.6.0")
 
     # This is an example of a dataset with multiple configurations.
     # If you don't want/need to define several sub-sets in your dataset,
@@ -83,26 +71,33 @@ class ArxivProofs(datasets.GeneratorBasedBuilder):
     BUILDER_CONFIGS = [
         datasets.BuilderConfig(name="proofs", version=VERSION, description="One proof per line"),
         datasets.BuilderConfig(name="sentences", version=VERSION, description="One sentence per line"),
+        datasets.BuilderConfig(name="raw", version=VERSION, description="One (less agressively cleaned) proof per line"),
+        datasets.BuilderConfig(name="tags", version=VERSION, description="arXiv subject tags for each paper"),
     ]
 
     DEFAULT_CONFIG_NAME = "proofs"  # It's not mandatory to have a default configuration. Just use one if it make sense.
 
     def _info(self):
         # TODO: This method specifies the datasets.DatasetInfo object which contains informations and typings for the dataset
-        if self.config.name == "proofs":  # This is the name of the configuration selected in BUILDER_CONFIGS above
+        if self.config.name in {"proofs", "raw"}:  # This is the name of the configuration selected in BUILDER_CONFIGS above
             features = datasets.Features(
                 {
-                    "fileID": datasets.Value("string"),
+                    "paper": datasets.Value("string"),
                     "proof": datasets.Value("string"),
-                    # These are the features of your dataset like images, labels ...
+                }
+            )
+        elif self.config.name == "tags":  # This is an example to show how to have different features for "proofs" and "sentences"
+            features = datasets.Features(
+                {
+                    "paper": datasets.Value("string"),
+                    "tags": datasets.Value("string"),
                 }
             )
         else:  # This is an example to show how to have different features for "proofs" and "sentences"
             features = datasets.Features(
                 {
-                    "fileID": datasets.Value("string"),
+                    "paper": datasets.Value("string"),
                     "sentence": datasets.Value("string"),
-                    # These are the features of your dataset like images, labels ...
                 }
             )
         return datasets.DatasetInfo(
@@ -136,8 +131,8 @@ class ArxivProofs(datasets.GeneratorBasedBuilder):
                 name=datasets.Split.TRAIN,
                 # These kwargs will be passed to _generate_examples
                 gen_kwargs={
-                    "filepath": data_dir, # os.path.join(data_dir, "train.jsonl"), # data_file,
-                    "split": "train",
+                    "filepath": data_dir,
+                    "split": "train",   # Prooflang doesn't have a train/test split.
                 },
             ),
             # datasets.SplitGenerator(
@@ -162,8 +157,9 @@ class ArxivProofs(datasets.GeneratorBasedBuilder):
     def _generate_examples(self, filepath, split):
         # TODO: This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
         # The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
+        csv.field_size_limit(256000)  # Some of the raw proofs are slightly longer than 131072 characters
         with open(os.path.join(filepath, self.config.name + ".tsv"), encoding="utf-8") as f:
-            reader = csv.DictReader(f, delimiter='\t')
+            reader = csv.DictReader(f, delimiter='\t', quoting=csv.QUOTE_NONE)
             for key, data in enumerate(reader):
                 yield key, data
                 # if self.config.name == "proofs":

proofs.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:a3cecb6a1fe54e2d80baafa727629786d64f7d212c08e80e2249753e81a90ccc
-size 723931097
+oid sha256:834fc89ed12a1e2f4ac3137195dd4c181f7c7083df9703c5069ee224c5aebe73
+size 722830298

raw.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:22564809efc16d396ea0470d848d9bc3f77ddcc6d2ddbc455cc2ed34c08fa51a
+size 753908748

sentences.zip

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:25bcbb4637bbe587be43dedf53d124e69e55e5b70aa8c47f63d4226082d36c95
-size 756959311
+oid sha256:cef9b89dbea688e2e7f7eb5f421f8ef5168e94c408a25701f2b4e22d8f70f205
+size 747935508

tags.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7cb505f1b7cc9f24801f1e0601b4065c6aeec7bfa35d3bf20477e17a74b60ae8
+size 1921707