warn-scraper 1.2.146.dev0__py3-none-any.whl → 1.2.149.dev0__py3-none-any.whl

warn/scrapers/mi.py

@@ -73,6 +73,7 @@ def scrape(
         "Total number of jobs impacted": "jobs",
         "Type of company action": "action",
         "Site address": "address",
+        "Site addresses": "address",
         "Reason for company action": "reason",
     }
 

warn/scrapers/nm.py

@@ -52,11 +52,15 @@ def scrape(
     document = BeautifulSoup(html, "html.parser")
     links = document.find_all("a")
     pdf_urls = [
-        f"{base_url}{link['href']}"
+        f"{link['href']}"
         for link in links
         if "WARN" in link.get("href", "") and link.get("href", "").endswith(".pdf")
     ]
 
+    for i, pdf_url in enumerate(pdf_urls):
+        if "http" not in pdf_url:
+            pdf_urls[i] = base_url + pdf_url
+
     output_rows = []
 
     for pdf_index, pdf_url in enumerate(pdf_urls):

warn/scrapers/ny.py

@@ -41,6 +41,14 @@ def scrape(
     Append .csv to the end of that URL:
     https://public.tableau.com/app/profile/kylee.teague2482/viz/WorkerAdjustmentRetrainingNotificationWARN/WARN.csv
     Try it in requests, no good. Try it in browser again. File downloads. Find it in the downloads section of the browser. Right-click, copy download link, try that in requests and ... it worked?
+
+    In 2026, the previous implementation broke. .csv appending gets a 403.
+    Wondering if we should be parsing this:
+    https://public.tableau.com/workbooks/WorkerAdjustmentRetrainingNotificationWARN.twb
+    using the Tableau API
+    https://github.com/tableau/document-api-python
+
+
     """
 
     url = "https://public.tableau.com/views/WorkerAdjustmentRetrainingNotificationWARN/WARN.csv?%3Adisplay_static_image=y&%3AbootstrapWhenNotified=true&%3Aembed=true&%3Alanguage=en-US&:embed=y&:showVizHome=n&:apiID=host0#navType=0&navSrc=Parse"

warn/scrapers/oh.py

@@ -49,6 +49,8 @@ def scrape(
     if isinstance(data_div, Tag):
         data = json.loads(data_div.decode_contents().strip())["data"]
     else:
+        logger.debug("!!!!!!!!!!!!!!!!!!!! Could not find JSON data div")
+        logger.debug(soup)
         raise ValueError("Could not find JSON data div")
     rawheaders = data[1]
     logger.debug(f"Found headers: {rawheaders}")

{warn_scraper-1.2.146.dev0.dist-info → warn_scraper-1.2.149.dev0.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: warn-scraper
-Version: 1.2.146.dev0
+Version: 1.2.149.dev0
 Summary: Command-line interface for downloading WARN Act notices of qualified plant closings and mass layoffs from state government websites
 Author-email: Big Local News <biglocalnews@stanford.edu>
 License-Expression: Apache-2.0

{warn_scraper-1.2.146.dev0.dist-info → warn_scraper-1.2.149.dev0.dist-info}/RECORD

@@ -30,14 +30,14 @@ warn/scrapers/ky.py,sha256=7kJTNOzxChyXlcyBImmdwwmrczYksU8XNxbhQ2owmJs,9688
 warn/scrapers/la.py,sha256=ORkMOQErl33SEiagOli4agDLdTt0R1MxxBmqOg3hNv8,13175
 warn/scrapers/md.py,sha256=hwgxXQnhyBWm8qF1dvxIThAX1MkrZbXLwRI9inO5t8g,4060
 warn/scrapers/me.py,sha256=q36F4yJ7hvZsLayA3uBS1romo4X3Qf-sEi2Y7LAQCi8,1172
-warn/scrapers/mi.py,sha256=APVKEIIcWqW0ul6LCS4QRk7ONJCv6voTsNovjdZX8so,5768
+warn/scrapers/mi.py,sha256=n-fgU44FkpFeWJ0Fis-pQR75AyFMGJqrEc93x2SIphc,5805
 warn/scrapers/mo.py,sha256=wnnwQAiVPwuheMqptMXZpyQdiKNghhKwTO-Bnh9oXoU,3492
 warn/scrapers/mt.py,sha256=t2MP4OCcuCEnrnvNgOu289P0eekZq4XaCK65qzgZX88,2457
 warn/scrapers/ne.py,sha256=JawuGJ3tCKvMd-N-p03gnltB4rol4QUJshMk2oyMPO4,4143
 warn/scrapers/nj.py,sha256=nwbMbeQuUJbYRVoyUyKZBmNqvqsXu3Habt-10r8DvZE,2230
-warn/scrapers/nm.py,sha256=HZpfLzn0LvLeRztYvqJ9n6FR5PYpyMndo8tzI8h9S2o,3581
-warn/scrapers/ny.py,sha256=hXbxPhiK-Eyc9h_05wkAsfdVIT0vayKX4EE5aiJVdBc,2291
-warn/scrapers/oh.py,sha256=6QD2Ymgqx-09OvY5bL-MGNnfmS-Y8CHQ00kLADSXcjE,3330
+warn/scrapers/nm.py,sha256=55kDh65uOc-uMO8wuxPHhHN4Aw8Di2Fx66JBds8f2Io,3694
+warn/scrapers/ny.py,sha256=_nYoawIde4yYtk8-tMG3hekLlSlrKKqULmkBWOUoQJM,2586
+warn/scrapers/oh.py,sha256=-_MyuaUGyKgd-XiXMVnf6AHkqOBgAFg0NRRpXGLz-3A,3431
 warn/scrapers/ok.py,sha256=ZZciyR1jPS4SzS2JSQwhJsDXP_VxA9UkEQvLpxzWzp4,7676
 warn/scrapers/or.py,sha256=0PjyrW3CHdxtHhqEo3Ob-9B6YckACoBD3K0c4FPQUcg,5208
 warn/scrapers/pa.py,sha256=j6cYZjKJUht2s829VPAbI2IRJazkqOwbwcVMAcItuIc,5410
@@ -51,9 +51,9 @@ warn/scrapers/va.py,sha256=7Nle7qL0VNPiE653XyaP9HQqSfuJFDRr2kEkjOqLvFM,11269
 warn/scrapers/vt.py,sha256=d-bo4WK2hkrk4BhCCmLpEovcoZltlvdIUB6O0uaMx5A,1186
 warn/scrapers/wa.py,sha256=UXdVtHZo_a-XfoiyOooTRfTb9W3PErSZdKca6SRORgs,4282
 warn/scrapers/wi.py,sha256=ClEzXkwZbop0W4fkQgsb5oHAPUrb4luUPGV-jOKwkcg,4855
-warn_scraper-1.2.146.dev0.dist-info/licenses/LICENSE,sha256=ZV-QHyqPwyMuwuj0lI05JeSjV1NyzVEk8Yeu7FPtYS0,585
-warn_scraper-1.2.146.dev0.dist-info/METADATA,sha256=c6QUDVzea1wxj2mkFmPv2f9PNUVMa7vHIFyoh4GwMmg,1780
-warn_scraper-1.2.146.dev0.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
-warn_scraper-1.2.146.dev0.dist-info/entry_points.txt,sha256=poh_oSweObGlBSs1_2qZmnTodlOYD0KfO7-h7W2UQIw,47
-warn_scraper-1.2.146.dev0.dist-info/top_level.txt,sha256=dZfms6N3kqVXufiPOo7YqOrAcUtYfNH_oyGvYUk9FB4,5
-warn_scraper-1.2.146.dev0.dist-info/RECORD,,
+warn_scraper-1.2.149.dev0.dist-info/licenses/LICENSE,sha256=ZV-QHyqPwyMuwuj0lI05JeSjV1NyzVEk8Yeu7FPtYS0,585
+warn_scraper-1.2.149.dev0.dist-info/METADATA,sha256=VQt1KDfh3oqeCaGriI-MZd5C_0QuXHnOTbCiYhQV42k,1780
+warn_scraper-1.2.149.dev0.dist-info/WHEEL,sha256=aeYiig01lYGDzBgS8HxWXOg3uV61G9ijOsup-k9o1sk,91
+warn_scraper-1.2.149.dev0.dist-info/entry_points.txt,sha256=poh_oSweObGlBSs1_2qZmnTodlOYD0KfO7-h7W2UQIw,47
+warn_scraper-1.2.149.dev0.dist-info/top_level.txt,sha256=dZfms6N3kqVXufiPOo7YqOrAcUtYfNH_oyGvYUk9FB4,5
+warn_scraper-1.2.149.dev0.dist-info/RECORD,,