Delete prepare_corpus.py

prepare_corpus.py

@@ -1,124 +0,0 @@
-import argparse
-import os
-import xmltodict
-import glob
-
-
-def extract_train(train_inp, train_out, docids_out=None):
-    with open(train_inp, "r", encoding="utf-8") as f:
-        all_docs = f.read()
-        docs_xml = [f"{d}</doc>" for d in all_docs.split("</doc>")[:-1]]
-        docs = [xmltodict.parse(doc) for doc in docs_xml]
-
-    train_out_f = open(train_out, "w", encoding="utf-8")
-    if docids_out is not None:
-        docids_out_f = open(docids_out, "w", encoding="utf-8")
-
-    for doc in docs:
-        if "#text" not in doc["doc"]:
-            continue
-        doc_id = doc["doc"]["@docid"]
-        for line in doc["doc"]["#text"].split("\n"):
-            print(line.strip(), file=train_out_f)
-            if docids_out is not None:
-                print(doc_id, file=docids_out_f)
-
-
-def extract_eval(eval_inp, eval_out, docids_out=None):
-    with open(eval_inp, "r", encoding="utf-8") as f:
-        all_docs = f.read()
-        docs_xml = xmltodict.parse(all_docs)
-        docs = (
-            docs_xml["mteval"]["srcset"]["doc"]
-            if "srcset" in docs_xml["mteval"]
-            else docs_xml["mteval"]["refset"]["doc"]
-        )
-
-    eval_out_f = open(eval_out, "a")
-    if docids_out is not None:
-        docids_out_f = open(docids_out, "a")
-
-    for doc in docs:
-        doc_id = doc["@docid"]
-        if not doc["seg"]:
-            continue
-        for seg in doc["seg"]:
-            line = seg["#text"]
-            print(line.strip(), file=eval_out_f)
-            if docids_out is not None:
-                print(doc_id, file=docids_out_f)
-
-
-if __name__ == "__main__":
-    parser = argparse.ArgumentParser("parser for iwslt2017 data")
-
-    parser.add_argument("raw_data")
-    parser.add_argument("out_data")
-    parser.add_argument("-s", "--source-lang", required=True, type=str)
-    parser.add_argument("-t", "--target-lang", required=True, type=str)
-    parser.add_argument(
-        "--devsets", nargs="*", default=["tst2011", "tst2012", "tst2013", "tst2014"]
-    )
-    parser.add_argument("--testsets", nargs="*", default=["tst2015"])
-    args = parser.parse_args()
-
-    first_l = (
-        args.source_lang
-        if glob.glob(f"{args.raw_data}/*.{args.source_lang}-{args.target_lang}.*")
-        else args.target_lang
-    )
-    second_l = (
-        args.target_lang
-        if glob.glob(f"{args.raw_data}/*.{args.source_lang}-{args.target_lang}.*")
-        else args.source_lang
-    )
-
-    train_inp_prefix = os.path.join(args.raw_data, f"train.tags.{first_l}-{second_l}")
-    train_out_prefix = os.path.join(
-        args.out_data, f"train.{args.source_lang}-{args.target_lang}"
-    )
-    extract_train(
-        f"{train_inp_prefix}.{args.source_lang}",
-        f"{train_out_prefix}.{args.source_lang}",
-        docids_out=f"{train_out_prefix}.docids",
-    )
-    extract_train(
-        f"{train_inp_prefix}.{args.target_lang}",
-        f"{train_out_prefix}.{args.target_lang}",
-    )
-
-    # generate valid set based on devsets paseed
-    valid_out_prefix = os.path.join(
-        args.out_data, f"valid.{args.source_lang}-{args.target_lang}"
-    )
-    for devset in args.devsets:
-        valid_inp_prefix = os.path.join(
-            args.raw_data, f"IWSLT17.TED.{devset}.{first_l}-{second_l}"
-        )
-        extract_eval(
-            f"{valid_inp_prefix}.{args.source_lang}.xml",
-            f"{valid_out_prefix}.{args.source_lang}",
-            docids_out=f"{valid_out_prefix}.docids",
-        )
-        extract_eval(
-            f"{valid_inp_prefix}.{args.target_lang}.xml",
-            f"{valid_out_prefix}.{args.target_lang}",
-        )
-
-    # generate test set based on testsets paseed
-    test_out_prefix = os.path.join(
-        args.out_data, f"test.{args.source_lang}-{args.target_lang}"
-    )
-    for testset in args.testsets:
-        test_inp_prefix = os.path.join(
-            args.raw_data, f"IWSLT17.TED.{testset}.{first_l}-{second_l}"
-        )
-        extract_eval(
-            f"{test_inp_prefix}.{args.source_lang}.xml",
-            f"{test_out_prefix}.{args.source_lang}",
-            docids_out=f"{test_out_prefix}.docids",
-        )
-        extract_eval(
-            f"{test_inp_prefix}.{args.target_lang}.xml",
-            f"{test_out_prefix}.{args.target_lang}",
-        )