rbiswasfc commited on
Commit
43bc104
1 Parent(s): 67be06c
Files changed (6) hide show
  1. Dockerfile +32 -0
  2. app.py +53 -0
  3. main.py +27 -2
  4. requirements.txt +13 -0
  5. supervisord.conf +20 -0
  6. zotero_refactor.ipynb +0 -746
Dockerfile CHANGED
@@ -0,0 +1,32 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ FROM python:3.10
2
+
3
+ RUN useradd -m -u 1000 user
4
+ USER user
5
+ ENV HOME=/home/user \
6
+ PATH=/home/user/.local/bin:$PATH
7
+
8
+ # Set the working directory
9
+ WORKDIR $HOME/app
10
+
11
+ COPY requirements.txt .
12
+ RUN pip install --no-cache-dir -r requirements.txt
13
+ RUN git config --global credential.helper store
14
+
15
+ COPY . .
16
+ COPY supervisord.conf .
17
+
18
+ # Set permissions on the log file
19
+ USER root
20
+ RUN touch $HOME/app/mylog.log $HOME/app/supervisord.log && chmod a+rwx $HOME/app/mylog.log $HOME/app/supervisord.log
21
+ USER user
22
+
23
+ RUN mkdir -p /tmp/cache/
24
+ RUN mkdir -p /.cache
25
+ RUN chmod a+rwx -R /tmp/cache/
26
+ RUN chmod a+rwx -R /.cache
27
+ ENV HF_HUB_CACHE=HF_HOME
28
+
29
+ ENV PYTHONUNBUFFERED=1 PORT=7860
30
+
31
+ # Run supervisord
32
+ CMD ["supervisord", "-c", "supervisord.conf"]
app.py CHANGED
@@ -0,0 +1,53 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import os
2
+
3
+ import dotenv
4
+ import pandas as pd
5
+ from datasets import load_dataset
6
+ from fasthtml.common import *
7
+
8
+ dotenv.load_dotenv()
9
+
10
+ app, rt = fast_app()
11
+ from huggingface_hub import login, whoami
12
+
13
+ login(token=os.environ.get("HF_TOKEN"))
14
+
15
+ hf_user = whoami(os.environ.get("HF_TOKEN"))["name"]
16
+ HF_REPO_ID = f"{hf_user}/zotero-answer-ai-articles"
17
+
18
+
19
+ @rt("/")
20
+ async def get():
21
+ """Dataset viewer home page"""
22
+ return Titled(
23
+ "Zotero Dataset Viewer",
24
+ Form(
25
+ Input(id="dataset_subset", placeholder="Enter dataset subset"),
26
+ Button("Load Dataset"),
27
+ hx_post="/load_dataset",
28
+ hx_target="#dataset_content",
29
+ ),
30
+ Div(id="dataset_content"),
31
+ )
32
+
33
+
34
+ @rt("/load_dataset")
35
+ async def post(dataset_subset: str):
36
+ """Load and display dataset"""
37
+ try:
38
+ dataset = load_dataset(HF_REPO_ID, dataset_subset, split="train")
39
+ df = pd.DataFrame(dataset[:10]) # Load first 10 rows
40
+ table = df.to_html(classes="table", index=False)
41
+
42
+ return Div(
43
+ H2(f"Dataset: {dataset_name}"),
44
+ P(f"Number of rows: {len(dataset)}"),
45
+ P(f"Columns: {', '.join(df.columns)}"),
46
+ Div(NotStr(table), cls="table-responsive"),
47
+ )
48
+
49
+ except Exception as e:
50
+ return Div(f"Error loading dataset: {str(e)}")
51
+
52
+
53
+ serve()
main.py CHANGED
@@ -1,9 +1,11 @@
1
  import os
2
  import re
 
3
 
4
  import dotenv
5
  import pandas as pd
6
  import requests
 
7
  import srsly
8
  from bs4 import BeautifulSoup
9
  from datasets import Dataset, load_dataset
@@ -372,8 +374,11 @@ def main():
372
  processed_arxiv_ids.add(item["arxiv_id"])
373
  except Exception as e:
374
  print(f"Failed to parse html for {item['arxiv_id']}: {e}")
 
 
 
375
  print("Extracting from pdf...")
376
- md_content = get_pdf_text(item["arxiv_id"])
377
  if md_content:
378
  item["contents"] = parse_markdown_content(md_content, item["arxiv_id"])
379
  processed_arxiv_ids.add(item["arxiv_id"])
@@ -431,5 +436,25 @@ def main():
431
  srsly.write_json("data/processed_arxiv_ids.json", processed_arxiv_ids)
432
 
433
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
434
  if __name__ == "__main__":
435
- main()
 
1
  import os
2
  import re
3
+ import time
4
 
5
  import dotenv
6
  import pandas as pd
7
  import requests
8
+ import schedule
9
  import srsly
10
  from bs4 import BeautifulSoup
11
  from datasets import Dataset, load_dataset
 
374
  processed_arxiv_ids.add(item["arxiv_id"])
375
  except Exception as e:
376
  print(f"Failed to parse html for {item['arxiv_id']}: {e}")
377
+ item["contents"] = []
378
+
379
+ if len(item["contents"]) == 0:
380
  print("Extracting from pdf...")
381
+ md_content = get_pdf_text(item["arxiv_id"]) # fix this
382
  if md_content:
383
  item["contents"] = parse_markdown_content(md_content, item["arxiv_id"])
384
  processed_arxiv_ids.add(item["arxiv_id"])
 
436
  srsly.write_json("data/processed_arxiv_ids.json", processed_arxiv_ids)
437
 
438
 
439
+ def schedule_periodic_task():
440
+ """
441
+ Schedule the main task to run at the user-defined frequency
442
+ """
443
+ main() # run once initially
444
+
445
+ frequency = "daily" # TODO: env
446
+ if frequency == "hourly":
447
+ print("Scheduling tasks to run every hour at the top of the hour")
448
+ schedule.every().hour.at(":00").do(main)
449
+ elif frequency == "daily":
450
+ start_time = "10:00"
451
+ print("Scheduling tasks to run every day at: {start_time} UTC+00")
452
+ schedule.every().day.at(start_time).do(main)
453
+
454
+ while True:
455
+ schedule.run_pending()
456
+ time.sleep(1)
457
+
458
+
459
  if __name__ == "__main__":
460
+ schedule_periodic_task()
requirements.txt CHANGED
@@ -0,0 +1,13 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ fasthtml-hf>=0.1.1
2
+ python-fasthtml>=0.0.8
3
+ huggingface-hub>=0.20.0
4
+ uvicorn>=0.29
5
+ schedule==1.2.0
6
+ supervisor==4.2.5
7
+ requests
8
+ srsly
9
+ python-dotenv
10
+ beautifulsoup4
11
+ retry
12
+ pandas
13
+ datasets
supervisord.conf ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [supervisord]
2
+ nodaemon=true
3
+
4
+ [program:main]
5
+ command=python main.py
6
+ stdout_logfile=/dev/stdout
7
+ stdout_logfile_maxbytes=0
8
+ stderr_logfile=/dev/stderr
9
+ stderr_logfile_maxbytes=0
10
+ autostart=true
11
+ # autorestart=true
12
+
13
+ [program:app]
14
+ command=python app.py
15
+ stdout_logfile=/dev/null
16
+ stdout_logfile_maxbytes=0
17
+ stderr_logfile=/dev/stderr
18
+ stderr_logfile_maxbytes=0
19
+ autostart=true
20
+ autorestart=true
zotero_refactor.ipynb DELETED
@@ -1,746 +0,0 @@
1
- {
2
- "cells": [
3
- {
4
- "cell_type": "code",
5
- "execution_count": 1,
6
- "id": "bfd057e4",
7
- "metadata": {},
8
- "outputs": [],
9
- "source": [
10
- "%ai reset"
11
- ]
12
- },
13
- {
14
- "cell_type": "code",
15
- "execution_count": 2,
16
- "id": "1c3aa2dc",
17
- "metadata": {},
18
- "outputs": [],
19
- "source": [
20
- "%%ai -c\n",
21
- "I'm refactoring the following code to ingest and process items from a zotero lib.\n",
22
- "\n",
23
- "```python\n",
24
- "import re\n",
25
- "\n",
26
- "import requests\n",
27
- "import srsly\n",
28
- "from bs4 import BeautifulSoup\n",
29
- "from retry import retry\n",
30
- "from tqdm.auto import tqdm\n",
31
- "\n",
32
- "\n",
33
- "\n",
34
- "def get_pdf_text(arxiv_id):\n",
35
- " url = \"http://147.189.194.113:80/extract\"\n",
36
- "\n",
37
- " try:\n",
38
- " response = requests.get(url, params={\"arxiv_id\": arxiv_id})\n",
39
- " response = response.json()\n",
40
- " if \"text\" in response:\n",
41
- " return response['text']\n",
42
- " return None\n",
43
- " except Exception as e:\n",
44
- " print(e)\n",
45
- " return None\n",
46
- " \n",
47
- "\n",
48
- "\n",
49
- "def get_zotero_items():\n",
50
- " \"\"\"fetch all items from zotero library\"\"\"\n",
51
- "\n",
52
- " GROUP_ID = \"5004697\"\n",
53
- " API_KEY = \"HFfTT8uuL7z9tN0L5VCHe7Ow\"\n",
54
- " BASE_URL = f\"https://api.zotero.org/groups/{GROUP_ID}/items\"\n",
55
- " LIMIT = 100\n",
56
- "\n",
57
- " headers = {\"Zotero-API-Key\": API_KEY, \"Content-Type\": \"application/json\"}\n",
58
- "\n",
59
- " items = []\n",
60
- " start = 0\n",
61
- "\n",
62
- " ITEM_TYPES = [\n",
63
- " \"journalArticle\",\n",
64
- " \"report\",\n",
65
- " \"conferencePaper\",\n",
66
- " \"manuscript\",\n",
67
- " \"book\",\n",
68
- " \"thesis\",\n",
69
- " ] # Replace or extend with desired item types\n",
70
- "\n",
71
- " i = 1\n",
72
- " while True:\n",
73
- " i += 1\n",
74
- " params = {\"limit\": LIMIT, \"start\": start}\n",
75
- " response = requests.get(BASE_URL, headers=headers, params=params)\n",
76
- " page_items = response.json()\n",
77
- " if ITEM_TYPES:\n",
78
- " params[\"itemType\"] = \",\".join(ITEM_TYPES)\n",
79
- "\n",
80
- " if not page_items:\n",
81
- " break\n",
82
- "\n",
83
- " items.extend(page_items)\n",
84
- " start += LIMIT\n",
85
- " print(f\"# items fetched {len(items)}\")\n",
86
- " # break # debug\n",
87
- "\n",
88
- " return items\n",
89
- "\n",
90
- "\n",
91
- "def extract_arxiv_links_and_pdfs(items):\n",
92
- " \"\"\"Extract mapping between arxiv & pdf urls for a list of zotero items\n",
93
- "\n",
94
- " Args:\n",
95
- " items (list): list of zotero items\n",
96
- "\n",
97
- " Returns:\n",
98
- " arxiv_info (dict): dictionary with arxiv links as keys and pdf links as values\n",
99
- " \"\"\"\n",
100
- " arxiv_info = {}\n",
101
- " arxiv_pattern = re.compile(r\"arxiv.org/abs/(\\d+\\.\\d+)\")\n",
102
- "\n",
103
- " for item in items:\n",
104
- " data = item.get(\"data\", {})\n",
105
- " attachments = item.get(\"links\", {}).get(\"attachment\", {})\n",
106
- " arxiv_url = None\n",
107
- " pdf_url = None\n",
108
- "\n",
109
- " if \"url\" in data and \"arxiv.org\" in data[\"url\"]:\n",
110
- " arxiv_match = arxiv_pattern.search(data[\"url\"])\n",
111
- " if arxiv_match:\n",
112
- " arxiv_url = data[\"url\"]\n",
113
- "\n",
114
- " if attachments:\n",
115
- " pdf_url = attachments[\"href\"]\n",
116
- "\n",
117
- " if arxiv_url:\n",
118
- " arxiv_info[arxiv_url] = pdf_url\n",
119
- "\n",
120
- " return arxiv_info\n",
121
- "\n",
122
- "\n",
123
- "@retry(tries=3, delay=15, backoff=2)\n",
124
- "def get_arxiv_with_retry(arxiv_id):\n",
125
- " return requests.get(f\"https://ar5iv.labs.arxiv.org/html/{arxiv_id}\")\n",
126
- "\n",
127
- "\n",
128
- "def get_arxiv(arxiv_ids):\n",
129
- " \"\"\"\n",
130
- " Fetch html version of arxiv papers\n",
131
- " \"\"\"\n",
132
- " html5ivs = {}\n",
133
- " for i, arxiv_id in tqdm(enumerate(arxiv_ids)):\n",
134
- " if \"v\" in arxiv_id:\n",
135
- " arxiv_id = arxiv_id.split(\"v\")[0]\n",
136
- " response = get_arxiv_with_retry(arxiv_id)\n",
137
- " html5ivs[arxiv_id] = response.text\n",
138
- " if response.status_code != 200:\n",
139
- " html5ivs[arxiv_id] = \"ERROR\"\n",
140
- " return html5ivs\n",
141
- "\n",
142
- "\n",
143
- "def extract_html_content(html):\n",
144
- " \"\"\"\n",
145
- " Extract content from arxiv html\n",
146
- " \"\"\"\n",
147
- " arxiv_id_match = re.search(r\"\\[(\\d+\\.\\d+(v\\d+)?)\\]\", html)\n",
148
- " arxiv_id = arxiv_id_match.group(1) if arxiv_id_match else None\n",
149
- " soup = BeautifulSoup(html, \"html.parser\")\n",
150
- " result = []\n",
151
- "\n",
152
- " # Extract paper title\n",
153
- " paper_title = soup.find(\"h1\", class_=\"ltx_title ltx_title_document\").get_text(strip=True)\n",
154
- "\n",
155
- " for math in soup.find_all(\"math\"):\n",
156
- " math.decompose()\n",
157
- " for cite in soup.find_all(\"cite\"):\n",
158
- " cite.decompose()\n",
159
- "\n",
160
- " # Extract abstract\n",
161
- " abstract = soup.find(\"div\", class_=\"ltx_abstract\")\n",
162
- " if abstract:\n",
163
- " result.append(\n",
164
- " {\n",
165
- " \"content\": \" \".join(p.get_text(strip=True) for p in abstract.find_all(\"p\")).replace(\")\", \") \"),\n",
166
- " \"title\": \"Abstract\",\n",
167
- " \"paper_title\": paper_title,\n",
168
- " \"content_type\": \"abstract\",\n",
169
- " }\n",
170
- " )\n",
171
- "\n",
172
- " # Extract introduction\n",
173
- " intro = soup.find(\"section\", id=\"S1\")\n",
174
- " if intro:\n",
175
- " section_title = \"Introduction\"\n",
176
- " paragraphs = intro.find_all(\"div\", class_=\"ltx_para\")\n",
177
- " for para in paragraphs:\n",
178
- " result.append(\n",
179
- " {\n",
180
- " \"content\": para.get_text(strip=True).replace(\")\", \") \"),\n",
181
- " \"title\": section_title,\n",
182
- " \"paper_title\": paper_title,\n",
183
- " \"content_type\": \"introduction\",\n",
184
- " }\n",
185
- " )\n",
186
- "\n",
187
- " # Extract other sections (excluding introduction and conclusion)\n",
188
- " sections = soup.find_all(\"section\", class_=\"ltx_section\")\n",
189
- " for section in sections[1:-1]:\n",
190
- " section_title = section.find(\"h2\", class_=\"ltx_title ltx_title_section\").get_text(strip=True)\n",
191
- " paragraphs = section.find_all(\"div\", class_=\"ltx_para\")\n",
192
- " for para in paragraphs:\n",
193
- " result.append(\n",
194
- " {\n",
195
- " \"content\": para.get_text(strip=True).replace(\")\", \") \"),\n",
196
- " \"title\": section_title,\n",
197
- " \"paper_title\": paper_title,\n",
198
- " \"content_type\": \"body\",\n",
199
- " }\n",
200
- " )\n",
201
- "\n",
202
- " # Extract conclusion\n",
203
- " conclusion = soup.find(\"section\", id=\"S4\")\n",
204
- " if conclusion:\n",
205
- " result.append(\n",
206
- " {\n",
207
- " \"content\": \" \".join(\n",
208
- " p.get_text(strip=True) for p in conclusion.find_all(\"div\", class_=\"ltx_para\")\n",
209
- " ).replace(\")\", \") \"),\n",
210
- " \"title\": conclusion.find(\"h2\", class_=\"ltx_title ltx_title_section\").get_text(strip=True),\n",
211
- " \"paper_title\": paper_title,\n",
212
- " \"content_type\": \"conclusion\",\n",
213
- " }\n",
214
- " )\n",
215
- "\n",
216
- " for c in result:\n",
217
- " c[\"arxiv_id\"] = arxiv_id\n",
218
- "\n",
219
- " return result\n",
220
- "\n",
221
- "\n",
222
- "def pool_html_content(parsed, pool_size=3, step_size=2):\n",
223
- " \"\"\"\n",
224
- " Pool together individual bits of contents from \"body\" and \"introduction\" sections where the \"title\" is the same.\n",
225
- " It pools them 'pool_size' by 'pool_size' as a sliding window with a step size of 'step_size'.\n",
226
- " Keeps all other attributes for each element and includes abstract and conclusion without pooling.\n",
227
- "\n",
228
- " Args:\n",
229
- " parsed (list): List of dictionaries containing parsed content.\n",
230
- " pool_size (int): Size of the pool for the sliding window.\n",
231
- " step_size (int): Step size for the sliding window.\n",
232
- "\n",
233
- " Returns:\n",
234
- " list: A list of pooled contents including non-pooled abstract and conclusion.\n",
235
- " \"\"\"\n",
236
- " pooled_contents = []\n",
237
- " i = 0\n",
238
- " while i < len(parsed) - pool_size + 1:\n",
239
- " current_window = parsed[i : i + pool_size]\n",
240
- " # Check if all entries in the current window have the same title and are from 'body' or 'introduction'\n",
241
- " if all(\n",
242
- " entry[\"content_type\"] in [\"body\", \"introduction\"] and entry[\"title\"] == current_window[0][\"title\"]\n",
243
- " for entry in current_window\n",
244
- " ):\n",
245
- " pooled_content = \"\\n\".join(entry[\"content\"] for entry in current_window)\n",
246
- " if \"bit\" not in current_window[0][\"title\"][:5]:\n",
247
- " title = current_window[0][\"title\"].lstrip(\"0123456789\")\n",
248
- " pooled_content = f\"{current_window[0]['paper_title']}\\n{title}\\n\\n\" + pooled_content\n",
249
- " pooled_entry = {\n",
250
- " \"title\": title,\n",
251
- " \"text\": pooled_content,\n",
252
- " \"paper_title\": current_window[0][\"paper_title\"],\n",
253
- " \"content_type\": current_window[0][\"content_type\"],\n",
254
- " \"arxiv_id\": current_window[0][\"arxiv_id\"],\n",
255
- " }\n",
256
- " pooled_contents.append(pooled_entry)\n",
257
- " i += step_size\n",
258
- "\n",
259
- " # Include abstract and conclusion without pooling\n",
260
- " for entry in parsed:\n",
261
- " if entry[\"content_type\"] in [\"abstract\", \"conclusion\"]:\n",
262
- " if \"content\" in entry:\n",
263
- " entry[\"text\"] = entry[\"content\"]\n",
264
- " del entry[\"content\"]\n",
265
- " if \"bit\" not in entry[\"title\"][:5]:\n",
266
- " entry[\"title\"] = entry[\"title\"].lstrip(\"0123456789\")\n",
267
- " entry[\"text\"] = f\"{entry['paper_title']}\\n{entry['title']}\\n\\n\" + entry[\"text\"]\n",
268
- " pooled_contents.append(entry)\n",
269
- "\n",
270
- " return pooled_contents\n",
271
- "\n",
272
- "\n",
273
- "def parse_markdown_content(md_content, arxiv_id):\n",
274
- " \"\"\"\n",
275
- " Parses markdown content to identify and extract sections based on headers.\n",
276
- " \"\"\"\n",
277
- "\n",
278
- " lines = md_content.split(\"\\n\")\n",
279
- " parsed = []\n",
280
- " current_section = None\n",
281
- " content = []\n",
282
- " paper_title = None\n",
283
- " current_title = None\n",
284
- "\n",
285
- " # Identify sections based on headers\n",
286
- " for line in lines:\n",
287
- " if line.startswith(\"#\"):\n",
288
- " if paper_title is None:\n",
289
- " paper_title = line.lstrip(\"#\").lstrip(\"#\").strip()\n",
290
- " continue\n",
291
- " if content:\n",
292
- " if not current_title:\n",
293
- " print(\"skipping pre-abstract!\")\n",
294
- " parsed.append(\n",
295
- " {\n",
296
- " \"title\": current_title,\n",
297
- " \"text\": content,\n",
298
- " \"paper_title\": paper_title,\n",
299
- " \"content_type\": \"body\"\n",
300
- " if current_section not in [\"Abstract\", \"Introduction\", \"Conclusion\"]\n",
301
- " else current_section.lower(),\n",
302
- " \"arxiv_id\": arxiv_id,\n",
303
- " }\n",
304
- " )\n",
305
- " content = []\n",
306
- "\n",
307
- " if \"Introduction\" in line:\n",
308
- " current_section = \"Introduction\"\n",
309
- " elif \"Abstract\" in line:\n",
310
- " current_section = \"Abstract\"\n",
311
- " elif \"Conclusion\" in line:\n",
312
- " current_section = \"Conclusion\"\n",
313
- " elif \"Reference\" in line:\n",
314
- " break\n",
315
- " else:\n",
316
- " current_section = \"body\"\n",
317
- "\n",
318
- " current_title = line.lstrip(\"#\").lstrip(\"#\").lstrip()\n",
319
- "\n",
320
- " if \"bit\" not in current_title:\n",
321
- " current_title = (\n",
322
- " current_title.lstrip(\"123456789\")\n",
323
- " .lstrip()\n",
324
- " .lstrip(\".\")\n",
325
- " .lstrip()\n",
326
- " .lstrip(\"123456789\")\n",
327
- " .lstrip()\n",
328
- " .lstrip(\".\")\n",
329
- " .lstrip()\n",
330
- " )\n",
331
- " else:\n",
332
- " content.append(line)\n",
333
- "\n",
334
- " if content:\n",
335
- " parsed.append(\n",
336
- " {\n",
337
- " \"title\": current_section,\n",
338
- " \"text\": content,\n",
339
- " \"paper_title\": paper_title,\n",
340
- " \"content_type\": \"body\"\n",
341
- " if current_section not in [\"Abstract\", \"Introduction\", \"Conclusion\"]\n",
342
- " else current_section.lower(),\n",
343
- " \"arxiv_id\": arxiv_id,\n",
344
- " }\n",
345
- " )\n",
346
- "\n",
347
- " return parsed\n",
348
- "\n",
349
- "\n",
350
- "def pool_markdown(parsed, pool_size=3, step_size=2):\n",
351
- " \"\"\"\n",
352
- " pool together markdown contents with sliding window of size 'pool_size' and step size of 'step_size'.\n",
353
- "\n",
354
- " Args:\n",
355
- " parsed (list): List of dictionaries containing parsed content.\n",
356
- " pool_size (int): Size of the pool for the sliding window.\n",
357
- " step_size (int): Step size for the sliding window.\n",
358
- "\n",
359
- " Returns:\n",
360
- " list: A list of pooled contents including non-pooled abstract and conclusion.\n",
361
- " \"\"\"\n",
362
- " cleaned_contents = []\n",
363
- " for section in parsed:\n",
364
- " cleaned_lines = [line for line in section[\"text\"] if line and \"|\" not in line and len(line) >= 40]\n",
365
- " section[\"text\"] = cleaned_lines\n",
366
- " cleaned_contents.append(section)\n",
367
- "\n",
368
- " pooled_contents = []\n",
369
- " for section in cleaned_contents:\n",
370
- " if section[\"content_type\"] in [\"body\", \"introduction\"]:\n",
371
- " i = 0\n",
372
- " while i <= len(section[\"text\"]) - pool_size:\n",
373
- " pooled_content = \"\\n\".join(section[\"text\"][i : i + pool_size])\n",
374
- " pooled_content = f\"{section['paper_title']}\\n{section['title']}\\n\\n\" + pooled_content\n",
375
- " pooled_entry = {\n",
376
- " \"title\": section[\"title\"],\n",
377
- " \"text\": pooled_content,\n",
378
- " \"paper_title\": section[\"paper_title\"],\n",
379
- " \"content_type\": section[\"content_type\"],\n",
380
- " \"arxiv_id\": section[\"arxiv_id\"],\n",
381
- " }\n",
382
- " pooled_contents.append(pooled_entry)\n",
383
- " i += step_size\n",
384
- " else:\n",
385
- " section[\"text\"] = f\"{section['paper_title']}\\n\" + \"\\n\".join(section[\"text\"])\n",
386
- " pooled_contents.append(section)\n",
387
- " return pooled_contents\n",
388
- "\n",
389
- "\n",
390
- "def clean_and_pool_markdown_content(md, arxiv_id):\n",
391
- " md = parse_markdown_content(md, arxiv_id)\n",
392
- " return pool_markdown(md)\n",
393
- "\n",
394
- "\n",
395
- "if __name__ == \"__main__\":\n",
396
- " items = get_zotero_items()\n",
397
- " print(f\"# of items fetched from zotero: {len(items)}\")\n",
398
- " arxiv_info = extract_arxiv_links_and_pdfs(items)\n",
399
- "\n",
400
- " # extract arxiv information ---\n",
401
- " arxiv_ids = set()\n",
402
- " for arxiv_url, pdf_url in arxiv_info.items():\n",
403
- " arxiv_ids.add(arxiv_url.split(\"/\")[-1])\n",
404
- " print(f\"# of arxiv papers: {len(arxiv_ids)}\")\n",
405
- "\n",
406
- " html5ivs = get_arxiv(arxiv_ids)\n",
407
- "\n",
408
- " broken = []\n",
409
- " extracted = {}\n",
410
- " extracted_from_pdf = {}\n",
411
- " broken_tex = {}\n",
412
- "\n",
413
- " pbar = tqdm(range(len(html5ivs)))\n",
414
- " for arxiv_id, x in html5ivs.items():\n",
415
- " try:\n",
416
- " extracted[arxiv_id] = extract_html_content(x)\n",
417
- " except Exception as e:\n",
418
- " # print(e)\n",
419
- " print(f'getting pdf text for {arxiv_id}...')\n",
420
- " pdf_text = get_pdf_text(arxiv_id) \n",
421
- " if pdf_text is not None:\n",
422
- " extracted_from_pdf[arxiv_id] = pdf_text # get_pdf_text(arxiv_id) # <- original\n",
423
- " else:\n",
424
- " broken.append(arxiv_id)\n",
425
- " pbar.update(1)\n",
426
- " pbar.close()\n",
427
- "\n",
428
- " print(f\"# of arxiv papers extracted: {len(extracted)}\")\n",
429
- " print(f\"# of arxiv papers extracted from pdf: {len(extracted_from_pdf)}\")\n",
430
- " print(f\"# of broken links: {len(broken)}\")\n",
431
- "\n",
432
- " pooled = [pool_html_content(x) for _, x in extracted.items()]\n",
433
- " pooled += [\n",
434
- " clean_and_pool_markdown_content(x, arxiv_id) for arxiv_id, x in extracted_from_pdf.items()\n",
435
- " ] # No pdf content yet, no op for now\n",
436
- " flat_pool = [item for sublist in pooled for item in sublist]\n",
437
- "\n",
438
- " id_to_abstract = {}\n",
439
- " for doc in pooled:\n",
440
- " for entry in doc:\n",
441
- " if entry[\"content_type\"] == \"abstract\":\n",
442
- " id_to_abstract[entry[\"arxiv_id\"]] = entry[\"text\"]\n",
443
- " break\n",
444
- "\n",
445
- " srsly.write_json(\"id_to_abstract.json\", id_to_abstract)\n",
446
- " srsly.write_json(\"flatpool.json\", flat_pool)\n",
447
- "```\n",
448
- "go through it and reply OK."
449
- ]
450
- },
451
- {
452
- "cell_type": "markdown",
453
- "id": "24e7ef23",
454
- "metadata": {},
455
- "source": [
456
- "OK. I've reviewed the code you provided for ingesting and processing items from a Zotero library."
457
- ]
458
- },
459
- {
460
- "cell_type": "markdown",
461
- "id": "84628bb2",
462
- "metadata": {},
463
- "source": [
464
- "OK. I've reviewed the code you provided for ingesting and processing items from a Zotero library."
465
- ]
466
- },
467
- {
468
- "cell_type": "code",
469
- "execution_count": 3,
470
- "id": "bada6931",
471
- "metadata": {},
472
- "outputs": [],
473
- "source": [
474
- "%%ai\n",
475
- "Let's work step by step. I currently have\n",
476
- "\n",
477
- "```\n",
478
- "import os\n",
479
- "import re\n",
480
- "\n",
481
- "import dotenv\n",
482
- "import requests\n",
483
- "from bs4 import BeautifulSoup\n",
484
- "from retry import retry\n",
485
- "from tqdm.auto import tqdm\n",
486
- "\n",
487
- "dotenv.load_dotenv()\n",
488
- "\n",
489
- "\n",
490
- "########################################################\n",
491
- "### GET ZOTERO ITEMS\n",
492
- "########################################################\n",
493
- "@retry(tries=3, delay=8)\n",
494
- "def _fetch_one_zotero_batch(url, headers, params):\n",
495
- " \"\"\"\n",
496
- " Fetch articles from Zotero API\n",
497
- " \"\"\"\n",
498
- " response = requests.get(url, headers=headers, params=params)\n",
499
- " response.raise_for_status()\n",
500
- " return response.json()\n",
501
- "\n",
502
- "\n",
503
- "def get_zotero_items(debug=False):\n",
504
- " \"\"\"\n",
505
- " fetch items from zotero library\n",
506
- " \"\"\"\n",
507
- "\n",
508
- " GROUP_ID = os.getenv(\"GROUP_ID\")\n",
509
- " API_KEY = os.getenv(\"API_KEY\")\n",
510
- " BASE_URL = f\"https://api.zotero.org/groups/{GROUP_ID}/items\"\n",
511
- " LIMIT = 100\n",
512
- "\n",
513
- " headers = {\"Zotero-API-Key\": API_KEY, \"Content-Type\": \"application/json\"}\n",
514
- "\n",
515
- " items = []\n",
516
- " start = 0\n",
517
- "\n",
518
- " i = 1\n",
519
- " while True:\n",
520
- " i += 1\n",
521
- " params = {\"limit\": LIMIT, \"start\": start}\n",
522
- " page_items = _fetch_one_zotero_batch(BASE_URL, headers, params)\n",
523
- "\n",
524
- " if not page_items:\n",
525
- " break\n",
526
- "\n",
527
- " items.extend(page_items)\n",
528
- " start += LIMIT\n",
529
- " print(f\"# items fetched {len(items)}\")\n",
530
- "\n",
531
- " if debug:\n",
532
- " break\n",
533
- "\n",
534
- " return items\n",
535
- "\n",
536
- "\n",
537
- "########################################################\n",
538
- "### EXTRACT ARXIV LINKS AND PDFs\n",
539
- "########################################################\n",
540
- "def get_arxiv_items(items):\n",
541
- " arxiv_items = []\n",
542
- " arxiv_pattern = re.compile(r\"arxiv.org/abs/(\\d+\\.\\d+)\")\n",
543
- "\n",
544
- " for item in items:\n",
545
- " data = item.get(\"data\", {})\n",
546
- " attachments = item.get(\"links\", {}).get(\"attachment\", {})\n",
547
- "\n",
548
- " arxiv_url = None\n",
549
- " pdf_url = None\n",
550
- "\n",
551
- " if \"url\" in data and \"arxiv.org\" in data[\"url\"]:\n",
552
- " arxiv_match = arxiv_pattern.search(data[\"url\"])\n",
553
- " if arxiv_match:\n",
554
- " arxiv_url = data[\"url\"]\n",
555
- "\n",
556
- " if attachments:\n",
557
- " pdf_url = attachments[\"href\"]\n",
558
- "\n",
559
- " if arxiv_url:\n",
560
- " arxiv_items.append(\n",
561
- " {\n",
562
- " \"arxiv_id\": arxiv_url.split(\"/\")[-1],\n",
563
- " \"arxiv_url\": arxiv_url,\n",
564
- " \"pdf_url\": pdf_url,\n",
565
- " \"added_by\": item[\"meta\"][\"createdByUser\"][\"username\"],\n",
566
- " \"date_added\": data.get(\"dateAdded\", \"\"),\n",
567
- " }\n",
568
- " )\n",
569
- "\n",
570
- " return arxiv_items\n",
571
- "\n",
572
- "\n",
573
- "@retry(tries=3, delay=15, backoff=2)\n",
574
- "def fetch_arxiv_html(arxiv_id):\n",
575
- " url = f\"https://ar5iv.labs.arxiv.org/html/{arxiv_id.split('v')[0]}\"\n",
576
- " response = requests.get(url)\n",
577
- " return response.text if response.status_code == 200 else None\n",
578
- "\n",
579
- "\n",
580
- "def get_arxiv_html(arxiv_items):\n",
581
- " for item in tqdm(arxiv_items):\n",
582
- " html = fetch_arxiv_html(item[\"arxiv_id\"])\n",
583
- " if html:\n",
584
- " item[\"raw_html\"] = html\n",
585
- " else:\n",
586
- " print(f\"failed to fetch html for {item['arxiv_id']}\")\n",
587
- "\n",
588
- " return arxiv_items\n",
589
- "\n",
590
- "\n",
591
- "if __name__ == \"__main__\":\n",
592
- " items = get_zotero_items(debug=True)\n",
593
- " print(f\"# of items fetched from zotero: {len(items)}\")\n",
594
- " arxiv_items = get_arxiv_items(items)\n",
595
- " print(f\"# of arxiv papers: {len(arxiv_items)}\")\n",
596
- " arxiv_htmls = get_arxiv_html(arxiv_items) # raw html files\n",
597
- "--\n",
598
- "review for any mistake until now. What shall we do next."
599
- ]
600
- },
601
- {
602
- "cell_type": "markdown",
603
- "id": "60ab1dfc",
604
- "metadata": {},
605
- "source": [
606
- "The code looks good so far. You've implemented functions to fetch items from Zotero, extract ArXiv-related items, and fetch the HTML content for ArXiv papers. \n",
607
- "\n",
608
- "Next steps could include:\n",
609
- "\n",
610
- "1. Parse the HTML content to extract structured information from the papers.\n",
611
- "2. Implement error handling for failed HTML fetches.\n",
612
- "3. Add a function to extract text from PDFs for papers where HTML fetch fails.\n",
613
- "4. Store the extracted data in a suitable format (e.g., JSON, database).\n",
614
- "5. Implement content pooling as in your original code.\n",
615
- "\n",
616
- "Which of these would you like to focus on first?"
617
- ]
618
- },
619
- {
620
- "cell_type": "code",
621
- "execution_count": 5,
622
- "id": "696318f6",
623
- "metadata": {},
624
- "outputs": [],
625
- "source": [
626
- "%%aip\n",
627
- "Let's work on step 1. Let's follow the original code as much as possible.\n",
628
- "Note: these are arxiv htmls. Let's also keep full body text."
629
- ]
630
- },
631
- {
632
- "cell_type": "code",
633
- "execution_count": 6,
634
- "id": "e7e4a5a6",
635
- "metadata": {},
636
- "outputs": [
637
- {
638
- "ename": "NameError",
639
- "evalue": "name 'arxiv_htmls' is not defined",
640
- "output_type": "error",
641
- "traceback": [
642
- "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
643
- "\u001b[0;31mNameError\u001b[0m Traceback (most recent call last)",
644
- "Cell \u001b[0;32mIn[6], line 18\u001b[0m\n\u001b[1;32m 15\u001b[0m \u001b[38;5;28;01mdef\u001b[39;00m \u001b[38;5;21mprocess_arxiv_items\u001b[39m(arxiv_items):\n\u001b[1;32m 16\u001b[0m \u001b[38;5;28;01mreturn\u001b[39;00m [\u001b[38;5;28mdict\u001b[39m(item, \u001b[38;5;241m*\u001b[39m\u001b[38;5;241m*\u001b[39mextract_html_content(item[\u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mraw_html\u001b[39m\u001b[38;5;124m\"\u001b[39m])) \u001b[38;5;28;01mfor\u001b[39;00m item \u001b[38;5;129;01min\u001b[39;00m arxiv_items \u001b[38;5;28;01mif\u001b[39;00m \u001b[38;5;124m\"\u001b[39m\u001b[38;5;124mraw_html\u001b[39m\u001b[38;5;124m\"\u001b[39m \u001b[38;5;129;01min\u001b[39;00m item]\n\u001b[0;32m---> 18\u001b[0m processed_items \u001b[38;5;241m=\u001b[39m process_arxiv_items(\u001b[43marxiv_htmls\u001b[49m)\n",
645
- "\u001b[0;31mNameError\u001b[0m: name 'arxiv_htmls' is not defined"
646
- ]
647
- }
648
- ],
649
- "source": [
650
- "def extract_html_content(html):\n",
651
- " soup = BeautifulSoup(html, \"html.parser\")\n",
652
- " arxiv_id = re.search(r\"\\[(\\d+\\.\\d+(v\\d+)?)\\]\", html).group(1) if re.search(r\"\\[(\\d+\\.\\d+(v\\d+)?)\\]\", html) else None\n",
653
- " paper_title = soup.find(\"h1\", class_=\"ltx_title ltx_title_document\").get_text(strip=True)\n",
654
- " [m.decompose() for m in soup.find_all([\"math\", \"cite\"])]\n",
655
- " abstract = \" \".join(p.get_text(strip=True) for p in soup.find(\"div\", class_=\"ltx_abstract\").find_all(\"p\")).replace(\")\", \") \") if soup.find(\"div\", class_=\"ltx_abstract\") else \"\"\n",
656
- " sections = soup.find_all(\"section\", class_=\"ltx_section\")\n",
657
- " body = []\n",
658
- " for section in sections:\n",
659
- " title = section.find(\"h2\", class_=\"ltx_title ltx_title_section\").get_text(strip=True)\n",
660
- " content = \" \".join(p.get_text(strip=True).replace(\")\", \") \") for p in section.find_all(\"div\", class_=\"ltx_para\"))\n",
661
- " body.append({\"title\": title, \"content\": content})\n",
662
- " return {\"arxiv_id\": arxiv_id, \"paper_title\": paper_title, \"abstract\": abstract, \"body\": body, \"full_text\": soup.get_text(strip=True)}\n",
663
- "\n",
664
- "def process_arxiv_items(arxiv_items):\n",
665
- " return [dict(item, **extract_html_content(item[\"raw_html\"])) for item in arxiv_items if \"raw_html\" in item]\n",
666
- "\n",
667
- "processed_items = process_arxiv_items(arxiv_htmls)"
668
- ]
669
- },
670
- {
671
- "cell_type": "code",
672
- "execution_count": 8,
673
- "id": "70a3caaa",
674
- "metadata": {},
675
- "outputs": [],
676
- "source": [
677
- "%%aip\n",
678
- "the above doesn't work and missed may details of the original implementation. Do not focus on pooling now\n",
679
- "just do extract_html_content better\n"
680
- ]
681
- },
682
- {
683
- "cell_type": "code",
684
- "execution_count": null,
685
- "id": "4040257d",
686
- "metadata": {},
687
- "outputs": [],
688
- "source": [
689
- "def extract_html_content(html):\n",
690
- " soup, result = BeautifulSoup(html, \"html.parser\"), []\n",
691
- " arxiv_id = re.search(r\"\\[(\\d+\\.\\d+(v\\d+)?)\\]\", html)\n",
692
- " arxiv_id = arxiv_id.group(1) if arxiv_id else None\n",
693
- " paper_title = soup.find(\"h1\", class_=\"ltx_title ltx_title_document\").get_text(strip=True)\n",
694
- " [m.decompose() for m in soup.find_all([\"math\", \"cite\"])]\n",
695
- " abstract = soup.find(\"div\", class_=\"ltx_abstract\")\n",
696
- " if abstract:\n",
697
- " result.append({\"content\": \" \".join(p.get_text(strip=True) for p in abstract.find_all(\"p\")).replace(\")\", \") \"), \"title\": \"Abstract\", \"paper_title\": paper_title, \"content_type\": \"abstract\"})\n",
698
- " intro = soup.find(\"section\", id=\"S1\")\n",
699
- " if intro:\n",
700
- " for para in intro.find_all(\"div\", class_=\"ltx_para\"):\n",
701
- " result.append({\"content\": para.get_text(strip=True).replace(\")\", \") \"), \"title\": \"Introduction\", \"paper_title\": paper_title, \"content_type\": \"introduction\"})\n",
702
- " for section in soup.find_all(\"section\", class_=\"ltx_section\")[1:-1]:\n",
703
- " section_title = section.find(\"h2\", class_=\"ltx_title ltx_title_section\").get_text(strip=True)\n",
704
- " for para in section.find_all(\"div\", class_=\"ltx_para\"):\n",
705
- " result.append({\"content\": para.get_text(strip=True).replace(\")\", \") \"), \"title\": section_title, \"paper_title\": paper_title, \"content_type\": \"body\"})\n",
706
- " conclusion = soup.find(\"section\", id=\"S4\")\n",
707
- " if conclusion:\n",
708
- " result.append({\"content\": \" \".join(p.get_text(strip=True) for p in conclusion.find_all(\"div\", class_=\"ltx_para\")).replace(\")\", \") \"), \"title\": conclusion.find(\"h2\", class_=\"ltx_title ltx_title_section\").get_text(strip=True), \"paper_title\": paper_title, \"content_type\": \"conclusion\"})\n",
709
- " for c in result:\n",
710
- " c[\"arxiv_id\"] = arxiv_id\n",
711
- " return result\n",
712
- "\n",
713
- "processed_items = [extract_html_content(item[\"raw_html\"]) for item in arxiv_items if \"raw_html\" in item]"
714
- ]
715
- },
716
- {
717
- "cell_type": "code",
718
- "execution_count": null,
719
- "id": "0242f021",
720
- "metadata": {},
721
- "outputs": [],
722
- "source": []
723
- }
724
- ],
725
- "metadata": {
726
- "kernelspec": {
727
- "display_name": "Python 3 (ipykernel)",
728
- "language": "python",
729
- "name": "python3"
730
- },
731
- "language_info": {
732
- "codemirror_mode": {
733
- "name": "ipython",
734
- "version": 3
735
- },
736
- "file_extension": ".py",
737
- "mimetype": "text/x-python",
738
- "name": "python",
739
- "nbconvert_exporter": "python",
740
- "pygments_lexer": "ipython3",
741
- "version": "3.12.4"
742
- }
743
- },
744
- "nbformat": 4,
745
- "nbformat_minor": 5
746
- }