antoinelb7 commited on
Commit
0faa90f
1 Parent(s): b81147e

Upload 12 files

Browse files
.gitattributes CHANGED
@@ -52,3 +52,8 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
52
  *.jpg filter=lfs diff=lfs merge=lfs -text
53
  *.jpeg filter=lfs diff=lfs merge=lfs -text
54
  *.webp filter=lfs diff=lfs merge=lfs -text
 
 
 
 
 
 
52
  *.jpg filter=lfs diff=lfs merge=lfs -text
53
  *.jpeg filter=lfs diff=lfs merge=lfs -text
54
  *.webp filter=lfs diff=lfs merge=lfs -text
55
+ data/alloprof.csv filter=lfs diff=lfs merge=lfs -text
56
+ data/pages/page-content-en.json filter=lfs diff=lfs merge=lfs -text
57
+ data/pages/page-content-fr.json filter=lfs diff=lfs merge=lfs -text
58
+ data/questions/comments.json filter=lfs diff=lfs merge=lfs -text
59
+ data/questions/discussions.json filter=lfs diff=lfs merge=lfs -text
data/alloprof.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:797a2cff4e85d0672918f17851358b68d3d2b15c54bf20f08260839b52c5d860
3
+ size 391561077
data/pages/page-content-en.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0e774b470f5ef18a48738bbecd75988cd409cb3b806524917a0fbae6fbc8313c
3
+ size 52195417
data/pages/page-content-fr.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:552ab80f3e31605fe233b418439cc7e1876c99e6fd1a3a7f8c4394115384b9b6
3
+ size 60883951
data/questions/categories.json ADDED
@@ -0,0 +1 @@
 
 
1
+ [{"CategoryID": -1, "Name": "Root"}, {"CategoryID": 2, "Name": "Anglais"}, {"CategoryID": 3, "Name": "Chimie"}, {"CategoryID": 4, "Name": "\u00c9ducation financi\u00e8re"}, {"CategoryID": 5, "Name": "Fran\u00e7ais"}, {"CategoryID": 6, "Name": "G\u00e9ographie"}, {"CategoryID": 7, "Name": "Histoire"}, {"CategoryID": 8, "Name": "Math\u00e9matiques"}, {"CategoryID": 9, "Name": "Monde contemporain"}, {"CategoryID": 10, "Name": "Physique"}, {"CategoryID": 11, "Name": "Sciences"}, {"CategoryID": 12, "Name": "Autre"}, {"CategoryID": 13, "Name": "Mathematics"}, {"CategoryID": 14, "Name": "English"}, {"CategoryID": 15, "Name": "History"}, {"CategoryID": 16, "Name": "Science"}, {"CategoryID": 17, "Name": "Physics"}, {"CategoryID": 18, "Name": "Chemistry"}, {"CategoryID": 19, "Name": "Geography"}, {"CategoryID": 20, "Name": "Contemporary World"}, {"CategoryID": 21, "Name": "Financial Education"}, {"CategoryID": 22, "Name": "French"}, {"CategoryID": 23, "Name": "Other"}]
data/questions/comments.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:380afab64d5a6152122faa2d77002d33bb853c46cfebe0a5e45362007e5de8f4
3
+ size 56383516
data/questions/discussions.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a1ca84b813c14a0f3f8d9e541cb2ca5cff43ca57a0ecaa3e30c1557e74fac46d
3
+ size 16264380
data/questions/grades.json ADDED
@@ -0,0 +1 @@
 
 
1
+ [{"GradeID": 0, "Name": "Utilisateur"}, {"GradeID": 1, "Name": "Primaire 1"}, {"GradeID": 2, "Name": "Primaire 2"}, {"GradeID": 3, "Name": "Primaire 3"}, {"GradeID": 4, "Name": "Primaire 4"}, {"GradeID": 5, "Name": "Primaire 5"}, {"GradeID": 6, "Name": "Primaire 6"}, {"GradeID": 7, "Name": "Secondaire 1"}, {"GradeID": 8, "Name": "Secondaire 2"}, {"GradeID": 9, "Name": "Secondaire 3"}, {"GradeID": 10, "Name": "Secondaire 4"}, {"GradeID": 11, "Name": "Secondaire 5"}, {"GradeID": 12, "Name": "Post-secondaire"}, {"GradeID": 13, "Name": "Parent"}, {"GradeID": 14, "Name": "Enseignant"}]
data/questions/roles.json ADDED
The diff for this file is too large to render. See raw diff
 
data/related_subjects.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "chemistry": ["other", "science", "physics"],
3
+ "contemporary_world": ["other", "history"],
4
+ "english": ["other"],
5
+ "financial_ed": ["other"],
6
+ "french": ["other"],
7
+ "geography": ["other"],
8
+ "history": ["other", "contemporary_world"],
9
+ "math": ["other", "physics"],
10
+ "other": [
11
+ "chemistry",
12
+ "contemporary_world",
13
+ "english",
14
+ "financial_ed",
15
+ "french",
16
+ "geography",
17
+ "history",
18
+ "math",
19
+ "physics",
20
+ "science"
21
+ ],
22
+ "physics": ["other", "science", "math", "chemistry"],
23
+ "science": ["other", "chemistry", "physics"]
24
+ }
scripts/download_images.py ADDED
@@ -0,0 +1,106 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/python
2
+ import json
3
+ import shutil
4
+ from pathlib import Path
5
+ from typing import Any, Iterable
6
+
7
+ import httpx
8
+ import polars as pl
9
+ import tqdm
10
+
11
+ ########
12
+ # main #
13
+ ########
14
+
15
+
16
+ def main() -> None:
17
+ path = Path(__file__).parent / ".." / "data"
18
+ load_print("Reading data...")
19
+ data = read_data(path)
20
+ urls = extract_image_urls(data)
21
+ done_print("Read data.")
22
+ load_print("Downloading images...")
23
+ urls = download_images(urls, path)
24
+ with open(path / "images" / "urls.json", "w") as f:
25
+ json.dump(urls, f, indent=2)
26
+ done_print("Downloaded images.")
27
+
28
+
29
+ def read_data(path: Path) -> pl.DataFrame:
30
+ return pl.read_csv(path / "alloprof.csv").with_columns(
31
+ [
32
+ pl.col("subject").str.split(";"),
33
+ pl.col("grade").str.split(";"),
34
+ pl.col("images").str.split(";"),
35
+ pl.col("relevant").str.split(";"),
36
+ pl.col("possible").str.split(";"),
37
+ ]
38
+ )
39
+
40
+
41
+ def extract_image_urls(data: pl.DataFrame) -> dict[str, int]:
42
+ return {
43
+ url: i
44
+ for i, url in enumerate(
45
+ set().union(*(set(row) for row in data["images"]))
46
+ )
47
+ }
48
+
49
+
50
+ def download_images(urls: dict[str, int], path: Path) -> dict[str, int]:
51
+ path = path / "images"
52
+ path.mkdir(exist_ok=True)
53
+ missing: list[str] = []
54
+ for url, id_ in load_progress(urls.items(), "Downloading images..."):
55
+ extension = url.split(".")[-1]
56
+ if extension in ("jpg", "jpeg", "png"):
57
+ with httpx.stream("GET", url) as resp:
58
+ if resp.status_code == 200:
59
+ with open(path / f"{id_}.{extension}", "wb") as f:
60
+ for chunk in resp.iter_bytes():
61
+ if chunk:
62
+ f.write(chunk)
63
+ else:
64
+ missing = [*missing, url]
65
+ else:
66
+ missing = [*missing, url]
67
+ return {url: id_ for url, id_ in urls.items() if url not in missing}
68
+
69
+
70
+ #########
71
+ # utils #
72
+ #########
73
+
74
+
75
+ def load_print(text: str, symbol: str = "*") -> None:
76
+ symbol = f"\033[1m[{symbol}]\033[0m"
77
+ print(
78
+ f"\r{symbol} {text}".ljust(shutil.get_terminal_size().columns),
79
+ end="\r",
80
+ )
81
+
82
+
83
+ def done_print(text: str, symbol: str = "+") -> None:
84
+ symbol = f"\033[1m\033[92m[{symbol}]\033[0m"
85
+ print(f"\r{symbol} {text}".ljust(shutil.get_terminal_size().columns))
86
+
87
+
88
+ def load_progress(
89
+ iter_: Iterable[Any],
90
+ text: str,
91
+ symbol: str = "*",
92
+ *args: Any,
93
+ **kwargs: Any,
94
+ ) -> Iterable[Any]:
95
+ symbol = f"\033[1m[{symbol}]\033[0m"
96
+ return tqdm.tqdm(
97
+ iter_,
98
+ f"\r{symbol} {text}",
99
+ *args,
100
+ leave=False,
101
+ **kwargs,
102
+ )
103
+
104
+
105
+ if __name__ == "__main__":
106
+ main()
scripts/parse_data.py ADDED
@@ -0,0 +1,566 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ #!/usr/bin/python
2
+ import argparse
3
+ import functools
4
+ import json
5
+ import re
6
+ import shutil
7
+ from pathlib import Path
8
+ from typing import Any, Iterator, cast
9
+
10
+ import polars as pl
11
+
12
+ ########
13
+ # main #
14
+ ########
15
+
16
+
17
+ def main() -> None:
18
+ args = read_args()
19
+
20
+ data = read_data(
21
+ args.related_subjects, args.grades_lower, args.grades_higher
22
+ )
23
+ done_print("Read data.")
24
+ load_print("Adding possible documents...")
25
+ data = add_possible_documents(data)
26
+ done_print("Added possible documents.")
27
+
28
+ load_print("Writing data...")
29
+ path = write_data(data)
30
+ done_print(
31
+ f"Wrote data to {path.resolve().relative_to(Path('.').resolve())}."
32
+ )
33
+ done_print("Data creation complete.")
34
+
35
+
36
+ def read_args() -> argparse.Namespace:
37
+ parser = argparse.ArgumentParser(
38
+ formatter_class=argparse.ArgumentDefaultsHelpFormatter
39
+ )
40
+ parser.add_argument(
41
+ "--related-subjects",
42
+ default="",
43
+ help="Path of the related subjects json file"
44
+ + " (see data/related_subjects.json)."
45
+ + " Leave empty to not use any related subjects.",
46
+ )
47
+ parser.add_argument(
48
+ "--grades-lower",
49
+ type=int,
50
+ default=0,
51
+ help="Number of grades lower than the minimum document grade for"
52
+ + " which to consider it possible.",
53
+ )
54
+ parser.add_argument(
55
+ "--grades-higher",
56
+ type=int,
57
+ default=0,
58
+ help="Number of grades higher than the maximum document grade for"
59
+ + " which to consider it possible.",
60
+ )
61
+ return parser.parse_args()
62
+
63
+
64
+ def read_data(
65
+ related_subjects_file: str, grades_lower: int, grades_higher: int
66
+ ) -> pl.DataFrame:
67
+ path = Path(__file__).parent / ".." / "data"
68
+ load_print("Reading questions...")
69
+ questions = read_questions(path)
70
+ done_print("Read questions.")
71
+ load_print("Reading pages...")
72
+ pages = read_pages(
73
+ path, related_subjects_file, grades_lower, grades_higher
74
+ )
75
+ done_print("Read pages.")
76
+ load_print("Combining pages and questions...")
77
+ return combine_documents(questions, pages)
78
+
79
+
80
+ def combine_documents(
81
+ questions: pl.DataFrame, pages: pl.DataFrame
82
+ ) -> pl.DataFrame:
83
+ # create hash of urls to quickly associate links to ids
84
+ page_hashes = {
85
+ hash(url): id_ for url, id_ in zip(pages["url"], pages["id"])
86
+ }
87
+
88
+ questions = (
89
+ questions.with_columns(
90
+ [
91
+ # convert all page links to the page ids, making sure each
92
+ # page can only appear once in the list
93
+ pl.col("page_links").apply(
94
+ lambda list_: list(
95
+ {
96
+ id_
97
+ for url in list_
98
+ if (
99
+ id_ := page_hashes.get(
100
+ hash(
101
+ url.replace(
102
+ "https://www.alloprof.qc.ca", ""
103
+ )
104
+ )
105
+ )
106
+ )
107
+ is not None
108
+ }
109
+ )
110
+ ),
111
+ # convert question links to its id
112
+ pl.col("question_links").apply(
113
+ lambda list_: list(
114
+ {
115
+ id_
116
+ for x in list_
117
+ if (
118
+ id_ := x.replace(
119
+ "https://www.alloprof.qc.ca/zonedentraide/discussion/", # noqa
120
+ "",
121
+ ).split("/")[0]
122
+ )
123
+ != "https:"
124
+ }
125
+ )
126
+ ),
127
+ ]
128
+ )
129
+ # combine page and question links in a single list of ids
130
+ .with_columns(
131
+ pl.col("page_links").arr.concat("question_links").alias("links")
132
+ )
133
+ .drop(["page_links", "question_links"])
134
+ .with_columns(
135
+ [
136
+ # to identify which document is a question or a page
137
+ pl.lit(True).alias("is_query"),
138
+ # to be able to find the actual question on the website
139
+ pl.col("id")
140
+ .apply(
141
+ lambda x: f"https://www.alloprof.qc.ca/zonedentraide/discussion/{x}" # noqa
142
+ )
143
+ .alias("url"),
144
+ ]
145
+ )
146
+ )
147
+
148
+ pages = (
149
+ pages
150
+ # add columns needed to concatenate to questions
151
+ .with_columns(
152
+ [
153
+ pl.col("id")
154
+ .apply(lambda _: [])
155
+ .cast(pl.List(pl.Utf8))
156
+ .alias("links"),
157
+ pl.lit(False).alias("is_query"),
158
+ pl.col("id")
159
+ .apply(lambda _: [])
160
+ .cast(pl.List(pl.Utf8))
161
+ .alias("images"),
162
+ pl.col("url").apply(
163
+ lambda x: f"https://www.alloprof.qc.ca{x}"
164
+ ),
165
+ ]
166
+ )
167
+ # have pages repeated for each grade and subject it's relevant for
168
+ .explode("grade").explode("subject")
169
+ )
170
+
171
+ return (
172
+ pl.concat([questions, pages], how="diagonal")
173
+ .rename({"links": "relevant"})
174
+ .with_columns(pl.col("relevant").apply(sorted))
175
+ )
176
+
177
+
178
+ def add_possible_documents(data: pl.DataFrame) -> pl.DataFrame:
179
+ # extract list of possible documents for each combination of subject,
180
+ # grade and language
181
+ data = data.with_columns(
182
+ (
183
+ pl.col("subject")
184
+ + ","
185
+ + pl.col("grade")
186
+ + ","
187
+ + pl.col("language")
188
+ ).alias("categories")
189
+ )
190
+ possible = (
191
+ data.select(["id", "categories"])
192
+ .unique()
193
+ .groupby("categories")
194
+ .agg(pl.list("id"))
195
+ .rename({"id": "possible"})
196
+ )
197
+ # add possible documents only to questions
198
+ data = pl.concat(
199
+ [
200
+ data.filter(pl.col("is_query"))
201
+ .join(possible, on="categories")
202
+ .drop("categories"),
203
+ data.filter(~pl.col("is_query")).with_columns(
204
+ pl.col("id")
205
+ .apply(lambda _: [])
206
+ .cast(pl.List(pl.Utf8))
207
+ .alias("possible")
208
+ ),
209
+ ],
210
+ how="diagonal",
211
+ )
212
+ # concatenate all subjects and grades for each document so there is only
213
+ # a single unique document per line
214
+ return (
215
+ # combine all grades for each id and subject
216
+ data.groupby(["id", "subject"])
217
+ .agg([pl.exclude("grade").first(), pl.list("grade")])
218
+ # combine all subjects for each id
219
+ .groupby("id")
220
+ .agg([pl.exclude("subject").first(), pl.list("subject")])
221
+ # remove duplicate grades
222
+ .with_columns(pl.col("grade").arr.unique())
223
+ )
224
+
225
+
226
+ def write_data(data: pl.DataFrame) -> Path:
227
+ path = Path(__file__).parent / ".." / "data" / "alloprof.csv"
228
+ data = data.with_columns(
229
+ [
230
+ pl.col("subject").arr.join(";"),
231
+ pl.col("grade").arr.join(";"),
232
+ pl.col("images").arr.join(";"),
233
+ pl.col("relevant").arr.join(";"),
234
+ pl.col("possible").arr.join(";"),
235
+ ]
236
+ )
237
+ data.write_csv(path)
238
+ return path
239
+
240
+
241
+ #############
242
+ # questions #
243
+ #############
244
+
245
+
246
+ def read_questions(path: Path) -> pl.DataFrame:
247
+ path = path / "questions"
248
+ questions = read_questions_(path)
249
+ answers = read_answers(path)
250
+ grades = read_grades(path)
251
+ subjects = read_subjects(path)
252
+
253
+ return (
254
+ questions
255
+ # convert subject and grade ids to their name
256
+ .join(subjects, on="CategoryID")
257
+ .drop("CategoryID")
258
+ .join(grades, on="GradeID")
259
+ .drop("GradeID")
260
+ # add answers and extract links and images
261
+ .join(answers, on="id", how="left")
262
+ .pipe(extract_relevant_links)
263
+ .with_columns(
264
+ [
265
+ pl.col("id").cast(pl.Utf8), # to make it consistent with pages
266
+ pl.col("text").apply(extract_text_from_json),
267
+ pl.col("text").apply(extract_images_from_json).alias("images"),
268
+ ]
269
+ )
270
+ )
271
+
272
+
273
+ def read_questions_(path: Path) -> pl.DataFrame:
274
+ return pl.read_json(path / "discussions.json").select(
275
+ [
276
+ pl.col("DiscussionID").alias("id"),
277
+ pl.col("Body").alias("text"),
278
+ pl.col("Language").alias("language"),
279
+ pl.col("InsertUserID").alias("user"),
280
+ pl.col("CategoryID"),
281
+ pl.col("GradeID"),
282
+ ]
283
+ )
284
+
285
+
286
+ def read_answers(path: Path) -> pl.DataFrame:
287
+ return (
288
+ pl.read_json(path / "comments.json")
289
+ .filter(~pl.col("DateAccepted").is_null())
290
+ .select(
291
+ [
292
+ pl.col("DiscussionID").alias("id"),
293
+ pl.col("Body").alias("answer"),
294
+ ]
295
+ )
296
+ )
297
+
298
+
299
+ def read_grades(path: Path) -> pl.DataFrame:
300
+ return pl.read_json(path / "grades.json").select(
301
+ [pl.col("GradeID"), pl.col("Name").alias("grade")]
302
+ )
303
+
304
+
305
+ def read_subjects(path: Path) -> pl.DataFrame:
306
+
307
+ return pl.read_json(path / "categories.json").select(
308
+ [
309
+ pl.col("CategoryID"),
310
+ # convert french subjects to english to make them consistent with
311
+ # pages
312
+ pl.col("Name").apply(convert_subject).alias("subject"),
313
+ ]
314
+ )
315
+
316
+
317
+ def extract_relevant_links(data: pl.DataFrame) -> pl.DataFrame:
318
+ def extract_links(text: str) -> list[str]:
319
+ return list(
320
+ set(
321
+ re.findall(
322
+ r"(https?:(?:\\)?/(?:\\)?/[a-zA-Z0-9/\\\.-]+)",
323
+ text.replace("\\/", "/"),
324
+ )
325
+ )
326
+ )
327
+
328
+ def extract_page_links(links: list[str]) -> list[str]:
329
+ return [link for link in links if "/eleves/bv/" in link]
330
+
331
+ def extract_question_links(links: list[str]) -> list[str]:
332
+ return [link for link in links if "/zonedentraide/discussion" in link]
333
+
334
+ return (
335
+ data.with_columns(pl.col("answer").fill_null(""))
336
+ .with_columns(pl.col("answer").apply(extract_links).alias("links"))
337
+ .with_columns(
338
+ [
339
+ pl.col("links").apply(extract_page_links).alias("page_links"),
340
+ pl.col("links")
341
+ .apply(extract_question_links)
342
+ .alias("question_links"),
343
+ ]
344
+ )
345
+ )
346
+
347
+
348
+ def extract_text_from_json(json_: str) -> str:
349
+
350
+ try:
351
+ return " ".join(list(extract_text(json.loads(json_))))
352
+ except json.JSONDecodeError:
353
+ return ""
354
+
355
+
356
+ def extract_text(raw_section: list[dict] | dict) -> Iterator[str]:
357
+ if isinstance(raw_section, list):
358
+ for section_content in raw_section:
359
+ yield from extract_text(section_content)
360
+
361
+ elif isinstance(raw_section, dict):
362
+ for section_tag, section_content in raw_section.items():
363
+ if section_tag == "insert" and isinstance(section_content, str):
364
+ yield re.sub(r"\s+", " ", section_content.strip())
365
+ elif section_tag == "url":
366
+ yield section_content.strip() # type: ignore
367
+ else:
368
+ yield from extract_text(section_content)
369
+
370
+
371
+ def extract_images_from_json(json_: str) -> list[str]:
372
+
373
+ try:
374
+ return list(extract_images(json.loads(json_)))
375
+ except json.JSONDecodeError:
376
+ return []
377
+
378
+
379
+ def extract_images(raw_section: list[dict] | dict) -> Iterator[str]:
380
+ if isinstance(raw_section, list):
381
+ for section_content in raw_section:
382
+ yield from extract_images(section_content)
383
+
384
+ elif isinstance(raw_section, dict):
385
+ for section_tag, section_content in raw_section.items():
386
+ if section_tag == "url":
387
+ yield cast(str, section_content)
388
+ else:
389
+ yield from extract_images(section_content)
390
+
391
+
392
+ #########
393
+ # pages #
394
+ #########
395
+
396
+
397
+ def read_pages(
398
+ path: Path,
399
+ related_subjects_file: str,
400
+ grades_lower: int,
401
+ grades_higher: int,
402
+ ) -> pl.DataFrame:
403
+ grades = read_grades(path / "questions")
404
+ fr_pages = pl.read_json(path / "pages" / "page-content-fr.json")["data"]
405
+ en_pages = pl.read_json(path / "pages" / "page-content-en.json")["data"]
406
+ return (
407
+ pl.DataFrame(
408
+ [parse_page_data(page) for page in [*fr_pages, *en_pages]]
409
+ )
410
+ .with_columns(
411
+ pl.col("subject")
412
+ .apply(convert_subject)
413
+ .apply(lambda subject: [subject])
414
+ )
415
+ .filter(pl.col("url") != "")
416
+ .pipe(
417
+ functools.partial(
418
+ convert_grades,
419
+ grades=grades,
420
+ grades_lower=grades_lower,
421
+ grades_higher=grades_higher,
422
+ )
423
+ )
424
+ .pipe(
425
+ functools.partial(
426
+ add_related_subjects,
427
+ related_subjects_file=related_subjects_file,
428
+ )
429
+ )
430
+ .pipe(extract_page_text)
431
+ )
432
+
433
+
434
+ def parse_page_data(data: dict[str, Any]) -> dict[str, str | int | list[str]]:
435
+ try:
436
+ page = {
437
+ "id": data["file"]["uuid"],
438
+ "url": data["file"]["breadcrumbs"]["current"]["routerLink"],
439
+ "language": data["file"]["lang"],
440
+ "subject": data["file"]["topic"],
441
+ "grade": data["file"]["levels"],
442
+ "title": data["file"]["title"],
443
+ "tags": data["file"]["tags"],
444
+ "content": " ".join(
445
+ d["attributes"]["content"]
446
+ for d in data["file"]["metatags"]
447
+ if d["attributes"]["content"]
448
+ ),
449
+ }
450
+ return {**page, "id": f"{page['id']}-{page['language']}"}
451
+ except TypeError:
452
+ return {}
453
+
454
+
455
+ def convert_grades(
456
+ data: pl.DataFrame,
457
+ grades: pl.DataFrame,
458
+ grades_lower: int,
459
+ grades_higher: int,
460
+ ) -> pl.DataFrame:
461
+ return (
462
+ # add grades lower and higher
463
+ data.with_columns(
464
+ pl.col("grade").apply(
465
+ lambda grades_: (
466
+ list(
467
+ range(
468
+ max(min(grades_) - grades_lower, 1),
469
+ min(grades_),
470
+ )
471
+ )
472
+ + list(grades_)
473
+ + list(
474
+ range(
475
+ max(grades_) + 1,
476
+ min(max(grades_) + grades_higher, 12) + 1,
477
+ )
478
+ )
479
+ )
480
+ if grades_ is not None
481
+ else []
482
+ )
483
+ )
484
+ # convert grades to their name
485
+ .with_columns(
486
+ pl.col("grade").apply(
487
+ lambda grades_: pl.DataFrame({"GradeID": grades_})
488
+ .join(grades, on="GradeID", how="left")["grade"]
489
+ .to_list()
490
+ )
491
+ )
492
+ )
493
+
494
+
495
+ def add_related_subjects(
496
+ data: pl.DataFrame, related_subjects_file: str
497
+ ) -> pl.DataFrame:
498
+ if related_subjects_file == "":
499
+ return data
500
+ else:
501
+ with open(related_subjects_file) as f:
502
+ related_subjects = json.load(f)
503
+ return data.with_columns(
504
+ pl.col("subject").apply(
505
+ lambda subject: list(subject) + related_subjects[subject[0]]
506
+ )
507
+ )
508
+
509
+
510
+ def extract_page_text(data: pl.DataFrame) -> pl.DataFrame:
511
+ return data.with_columns(
512
+ (
513
+ pl.col("title")
514
+ + " "
515
+ + pl.col("tags").arr.join(" ")
516
+ + " "
517
+ + pl.col("content")
518
+ ).alias("text")
519
+ ).drop(["title", "tags", "content"])
520
+
521
+
522
+ #########
523
+ # utils #
524
+ #########
525
+
526
+
527
+ def load_print(text: str, symbol: str = "*") -> None:
528
+ symbol = f"\033[1m[{symbol}]\033[0m"
529
+ print(
530
+ f"\r{symbol} {text}".ljust(shutil.get_terminal_size().columns),
531
+ end="\r",
532
+ )
533
+
534
+
535
+ def done_print(text: str, symbol: str = "+") -> None:
536
+ symbol = f"\033[1m\033[92m[{symbol}]\033[0m"
537
+ print(f"\r{symbol} {text}".ljust(shutil.get_terminal_size().columns))
538
+
539
+
540
+ def convert_subject(subject: str) -> str:
541
+ subject_conversions = {
542
+ "chemistry": ["chimie"],
543
+ "contemporary_world": ["monde contemporain", "contemporary world"],
544
+ "english": ["anglais"],
545
+ "financial_ed": ["éducation financière", "financial education"],
546
+ "french": ["français"],
547
+ "geography": ["géographie"],
548
+ "history": ["histoire"],
549
+ "math": ["mathématiques", "mathematics"],
550
+ "other": ["autre"],
551
+ "physics": ["physique"],
552
+ "science": ["sciences"],
553
+ }
554
+ match = [
555
+ key
556
+ for key, val in subject_conversions.items()
557
+ if subject.lower() in [key, *val]
558
+ ]
559
+ if match:
560
+ return match[0]
561
+ else:
562
+ return "other"
563
+
564
+
565
+ if __name__ == "__main__":
566
+ main()
scripts/requirements.txt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ httpx
2
+ polars
3
+ tqdm