Commit
•
52edc12
1
Parent(s):
8a5bfcb
Refactor
Browse files
alt.py
CHANGED
@@ -216,10 +216,8 @@ class Alt(datasets.GeneratorBasedBuilder):
|
|
216 |
|
217 |
data = {}
|
218 |
if self.config.name.startswith("alt-parallel"):
|
219 |
-
files = self.config.languages
|
220 |
-
|
221 |
data = {}
|
222 |
-
for lang in
|
223 |
file_path = os.path.join(basepath, "ALT-Parallel-Corpus-20191206", f"data_{lang}.txt")
|
224 |
with open(file_path, encoding="utf-8") as fin:
|
225 |
for line in fin:
|
@@ -227,16 +225,18 @@ class Alt(datasets.GeneratorBasedBuilder):
|
|
227 |
sp = line.split("\t")
|
228 |
|
229 |
_, urlid, sntid = sp[0].split(".")
|
230 |
-
|
|
|
231 |
if urlid not in allow_urls:
|
232 |
continue
|
233 |
|
234 |
if sntid not in data:
|
235 |
-
data[sntid] = {
|
236 |
-
|
237 |
-
|
238 |
-
|
239 |
-
|
|
|
240 |
|
241 |
# Note that Japanese and Myanmar texts have empty sentence fields in this release.
|
242 |
if len(sp) >= 2:
|
@@ -259,15 +259,10 @@ class Alt(datasets.GeneratorBasedBuilder):
|
|
259 |
"SNT.URLID": urlid,
|
260 |
"SNT.URLID.SNTID": sntid,
|
261 |
"url": allow_urls[urlid]["url"],
|
262 |
-
"status":
|
263 |
"value": sp[1],
|
264 |
}
|
265 |
|
266 |
-
if fname == "English-ALT-Draft.txt":
|
267 |
-
d["status"] = "draft"
|
268 |
-
else:
|
269 |
-
d["status"] = "reviewed"
|
270 |
-
|
271 |
data[sntid] = d
|
272 |
|
273 |
elif self.config.name == "alt-jp":
|
@@ -287,16 +282,12 @@ class Alt(datasets.GeneratorBasedBuilder):
|
|
287 |
"SNT.URLID.SNTID": sntid,
|
288 |
"url": allow_urls[urlid]["url"],
|
289 |
"value": sp[1],
|
290 |
-
|
291 |
"word_alignment": None,
|
292 |
"en_tokenized": None,
|
293 |
"jp_tokenized": None,
|
294 |
}
|
295 |
|
296 |
-
if fname == "Japanese-ALT-Draft.txt":
|
297 |
-
d["status"] = "draft"
|
298 |
-
else:
|
299 |
-
d["status"] = "reviewed"
|
300 |
data[sntid] = d
|
301 |
|
302 |
keys = {
|
@@ -354,9 +345,7 @@ class Alt(datasets.GeneratorBasedBuilder):
|
|
354 |
continue
|
355 |
|
356 |
k = "km_pos_tag" if fname == "data_km.km-tag.nova" else "km_tokenized"
|
357 |
-
if sntid in data:
|
358 |
-
data[sntid][k] = sp[1]
|
359 |
-
else:
|
360 |
data[sntid] = {
|
361 |
"SNT.URLID": urlid,
|
362 |
"SNT.URLID.SNTID": sntid,
|
@@ -364,7 +353,7 @@ class Alt(datasets.GeneratorBasedBuilder):
|
|
364 |
"km_pos_tag": None,
|
365 |
"km_tokenized": None,
|
366 |
}
|
367 |
-
|
368 |
|
369 |
elif self.config.name == "alt-my-transliteration":
|
370 |
file_path = os.path.join(basepath, "my-en-transliteration", "data.txt")
|
|
|
216 |
|
217 |
data = {}
|
218 |
if self.config.name.startswith("alt-parallel"):
|
|
|
|
|
219 |
data = {}
|
220 |
+
for lang in self.config.languages:
|
221 |
file_path = os.path.join(basepath, "ALT-Parallel-Corpus-20191206", f"data_{lang}.txt")
|
222 |
with open(file_path, encoding="utf-8") as fin:
|
223 |
for line in fin:
|
|
|
225 |
sp = line.split("\t")
|
226 |
|
227 |
_, urlid, sntid = sp[0].split(".")
|
228 |
+
# Some lines have a trailing blank space: "SNT.102053.5598 " in data_fil.txt
|
229 |
+
sntid = sntid.strip()
|
230 |
if urlid not in allow_urls:
|
231 |
continue
|
232 |
|
233 |
if sntid not in data:
|
234 |
+
data[sntid] = {
|
235 |
+
"SNT.URLID": urlid,
|
236 |
+
"SNT.URLID.SNTID": sntid,
|
237 |
+
"url": allow_urls[urlid]["url"],
|
238 |
+
"translation": {},
|
239 |
+
}
|
240 |
|
241 |
# Note that Japanese and Myanmar texts have empty sentence fields in this release.
|
242 |
if len(sp) >= 2:
|
|
|
259 |
"SNT.URLID": urlid,
|
260 |
"SNT.URLID.SNTID": sntid,
|
261 |
"url": allow_urls[urlid]["url"],
|
262 |
+
"status": "draft" if fname == "English-ALT-Draft.txt" else "reviewed",
|
263 |
"value": sp[1],
|
264 |
}
|
265 |
|
|
|
|
|
|
|
|
|
|
|
266 |
data[sntid] = d
|
267 |
|
268 |
elif self.config.name == "alt-jp":
|
|
|
282 |
"SNT.URLID.SNTID": sntid,
|
283 |
"url": allow_urls[urlid]["url"],
|
284 |
"value": sp[1],
|
285 |
+
"status": "draft" if fname == "Japanese-ALT-Draft.txt" else "reviewed",
|
286 |
"word_alignment": None,
|
287 |
"en_tokenized": None,
|
288 |
"jp_tokenized": None,
|
289 |
}
|
290 |
|
|
|
|
|
|
|
|
|
291 |
data[sntid] = d
|
292 |
|
293 |
keys = {
|
|
|
345 |
continue
|
346 |
|
347 |
k = "km_pos_tag" if fname == "data_km.km-tag.nova" else "km_tokenized"
|
348 |
+
if sntid not in data:
|
|
|
|
|
349 |
data[sntid] = {
|
350 |
"SNT.URLID": urlid,
|
351 |
"SNT.URLID.SNTID": sntid,
|
|
|
353 |
"km_pos_tag": None,
|
354 |
"km_tokenized": None,
|
355 |
}
|
356 |
+
data[sntid][k] = sp[1]
|
357 |
|
358 |
elif self.config.name == "alt-my-transliteration":
|
359 |
file_path = os.path.join(basepath, "my-en-transliteration", "data.txt")
|