warn-scraper 1.2.162.dev0__py3-none-any.whl → 1.2.164.dev0__py3-none-any.whl

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
warn/scrapers/tx.py CHANGED
@@ -1,8 +1,11 @@
1
1
  import logging
2
+ import random
2
3
  import re
3
4
  import sys
4
5
  from pathlib import Path
6
+ from time import sleep
5
7
 
8
+ import niquests as requests
6
9
  from bs4 import BeautifulSoup
7
10
  from openpyxl import load_workbook
8
11
 
@@ -32,22 +35,23 @@ def scrape(
32
35
 
33
36
  Returns: the Path where the file is written
34
37
  """
35
- ssl_verify = (
36
- True # Problems with certificates in November 2024; re-enabled March 2026
37
- )
38
-
39
38
  # Set up the cache
40
39
  cache = Cache(cache_dir)
41
40
 
42
41
  # Get the root URL
43
42
  url = "https://www.twc.texas.gov/data-reports/warn-notice"
44
43
  # pagebin, html = utils.get_with_zyte(url)
45
- page = utils.get_url(
46
- url,
47
- verify=ssl_verify,
48
- user_agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/999.0.0.0 Safari/537.36",
49
- )
44
+ session = requests.Session()
45
+
46
+ page = session.get(url)
47
+
48
+ # page = utils.get_url(
49
+ # url,
50
+ # verify=ssl_verify,
51
+ # user_agent="Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/999.0.0.0 Safari/537.36",
52
+ # )
50
53
  html = page.text
54
+ sleep(random.uniform(2, 4))
51
55
 
52
56
  # Cache it
53
57
  cache.write("tx/source.html", html)
@@ -86,7 +90,14 @@ def scrape(
86
90
  # download the excel file
87
91
  year = _get_year(href)
88
92
  ext = _get_ext(href)
89
- excel_path = cache.download(f"tx/{year}{ext}", data_url, verify=ssl_verify)
93
+ # excel_path = cache.download(f"tx/{year}{ext}", data_url, verify=ssl_verify)
94
+ r = session.get(data_url)
95
+ excel_path = cache_dir / f"tx/{year}{ext}"
96
+
97
+ with open(excel_path, "wb") as outfile:
98
+ outfile.write(r.content)
99
+ sleep(random.uniform(2, 4))
100
+
90
101
  # filename = f"tx/{year}{ext}"
91
102
  # excelbin, exceltext = utils.get_with_zyte(data_url)
92
103
  # excel_path = cache.write_binary(filename, excelbin)
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.4
2
2
  Name: warn-scraper
3
- Version: 1.2.162.dev0
3
+ Version: 1.2.164.dev0
4
4
  Summary: Command-line interface for downloading WARN Act notices of qualified plant closings and mass layoffs from state government websites
5
5
  Author-email: Big Local News <biglocalnews@stanford.edu>
6
6
  License-Expression: Apache-2.0
@@ -48,17 +48,17 @@ warn/scrapers/ri.py,sha256=EUyLy59eNiYHqiJR8C0YcJrZtp09KyVc45AFD0_Uc0U,4497
48
48
  warn/scrapers/sc.py,sha256=p3kscSNSW9C8C5QaSUbCAo6XibgB7G2iH6zaMH7Mnsc,4819
49
49
  warn/scrapers/sd.py,sha256=_4R19Ybzsyx1PvcWV3_laJmJ3etrwVGfhNEQm6njwoA,1904
50
50
  warn/scrapers/tn.py,sha256=i1H7c09Ea3CDrTXqqRMLBMPT_34QtGA0-x7T8rm_j5Q,2945
51
- warn/scrapers/tx.py,sha256=fmmQoi8BEWfZnKS9IFtlctztwPX_KDqNEkzaLyrr9fo,5301
51
+ warn/scrapers/tx.py,sha256=y_pRNuTzaQBkHpRKJXA5Fw8ZsjyPJIVElcwGinlf9wk,5594
52
52
  warn/scrapers/ut.py,sha256=iUh38YIjbvv5MyyKacsiZNe8KjfdBeDaOf-qMQEF_kc,2245
53
53
  warn/scrapers/va.py,sha256=7Nle7qL0VNPiE653XyaP9HQqSfuJFDRr2kEkjOqLvFM,11269
54
54
  warn/scrapers/vt.py,sha256=d-bo4WK2hkrk4BhCCmLpEovcoZltlvdIUB6O0uaMx5A,1186
55
55
  warn/scrapers/wa.py,sha256=UXdVtHZo_a-XfoiyOooTRfTb9W3PErSZdKca6SRORgs,4282
56
56
  warn/scrapers/wi.py,sha256=ClEzXkwZbop0W4fkQgsb5oHAPUrb4luUPGV-jOKwkcg,4855
57
- warn_scraper-1.2.162.dev0.dist-info/licenses/LICENSE,sha256=ZV-QHyqPwyMuwuj0lI05JeSjV1NyzVEk8Yeu7FPtYS0,585
58
- warn_scraper-1.2.162.dev0.dist-info/METADATA,sha256=iIxHyYYoXhKz95ChV9tTyJus6FK81vWODQf1GDCC_sc,1780
59
- warn_scraper-1.2.162.dev0.dist-info/WHEEL,sha256=K260EYznzXsJYBQGqmI8VTxEdiZYNvDZwW9cBh9-_MA,91
60
- warn_scraper-1.2.162.dev0.dist-info/entry_points.txt,sha256=poh_oSweObGlBSs1_2qZmnTodlOYD0KfO7-h7W2UQIw,47
61
- warn_scraper-1.2.162.dev0.dist-info/scm_file_list.json,sha256=SBROd7oEfcspIMl6v7yfGzQd2rkWxybkiIXopuTQVL0,4120
62
- warn_scraper-1.2.162.dev0.dist-info/scm_version.json,sha256=xg0s6KXFmkHX7NjoJHIc4J97T58iHzVCQ5h4vtd09qU,161
63
- warn_scraper-1.2.162.dev0.dist-info/top_level.txt,sha256=dZfms6N3kqVXufiPOo7YqOrAcUtYfNH_oyGvYUk9FB4,5
64
- warn_scraper-1.2.162.dev0.dist-info/RECORD,,
57
+ warn_scraper-1.2.164.dev0.dist-info/licenses/LICENSE,sha256=ZV-QHyqPwyMuwuj0lI05JeSjV1NyzVEk8Yeu7FPtYS0,585
58
+ warn_scraper-1.2.164.dev0.dist-info/METADATA,sha256=9uVo6pFhNTkENUnJQlk0sAEmSydHUtpIqgW0YYy8XoA,1780
59
+ warn_scraper-1.2.164.dev0.dist-info/WHEEL,sha256=K260EYznzXsJYBQGqmI8VTxEdiZYNvDZwW9cBh9-_MA,91
60
+ warn_scraper-1.2.164.dev0.dist-info/entry_points.txt,sha256=poh_oSweObGlBSs1_2qZmnTodlOYD0KfO7-h7W2UQIw,47
61
+ warn_scraper-1.2.164.dev0.dist-info/scm_file_list.json,sha256=SBROd7oEfcspIMl6v7yfGzQd2rkWxybkiIXopuTQVL0,4120
62
+ warn_scraper-1.2.164.dev0.dist-info/scm_version.json,sha256=o9jp_dH_h9sOfOdsitbgUoIVY45dh_C8KFlgYCGZeqI,161
63
+ warn_scraper-1.2.164.dev0.dist-info/top_level.txt,sha256=dZfms6N3kqVXufiPOo7YqOrAcUtYfNH_oyGvYUk9FB4,5
64
+ warn_scraper-1.2.164.dev0.dist-info/RECORD,,
@@ -1,7 +1,7 @@
1
1
  {
2
- "tag": "1.2.161",
2
+ "tag": "1.2.163",
3
3
  "distance": 0,
4
- "node": "g8cb14de08003c73f93a7388619e1f3ae02c41fd2",
4
+ "node": "g15659c6a544f7b833d5f3a979b1d03d6c3be9b1a",
5
5
  "dirty": true,
6
6
  "branch": "HEAD",
7
7
  "node_date": "2026-07-08"