parsagon 0.17.1__py3-none-any.whl → 0.17.2__py3-none-any.whl

parsagon/executor.py

@@ -16,6 +16,7 @@ import httpx
 from lxml import etree
 import lxml.html
 from lxml.html.clean import Cleaner
+import pymupdf4llm
 from pypdf import PdfReader
 from pyvirtualdisplay import Display
 import undetected_chromedriver as uc
@@ -706,6 +707,15 @@ class Executor:
         os.remove(most_recent_file)
         return text
 
+    def get_pdf_md(self, url):
+        window_id = self.goto(url)
+        self.close_window(window_id)
+        files = glob.glob("*")
+        most_recent_file = max(files, key=os.path.getmtime)
+        md_text = pymupdf4llm.to_markdown(most_recent_file)
+        os.remove(most_recent_file)
+        return md_text
+
     def str_to_iso8601(self, s):
         while s:
             try:

{parsagon-0.17.1.dist-info → parsagon-0.17.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: parsagon
-Version: 0.17.1
+Version: 0.17.2
 Summary: Allows you to create browser automations with natural language
 Author-email: Sandy Suh <sandy@parsagon.io>
 Project-URL: Homepage, https://parsagon.io
@@ -20,6 +20,8 @@ Requires-Dist: undetected-chromedriver==3.5.5
 Requires-Dist: webdriver-manager==4.0.2
 Requires-Dist: jsonpath-ng==1.5.3
 Requires-Dist: simplejson==3.19.1
+Requires-Dist: PyMuPDF==1.26.4
+Requires-Dist: pymupdf4llm==0.0.27
 Requires-Dist: pypdf==5.9.0
 Requires-Dist: python-dateutil==2.9.0.post0
 Requires-Dist: html2text==2024.2.26

{parsagon-0.17.1.dist-info → parsagon-0.17.2.dist-info}/RECORD

@@ -5,7 +5,7 @@ parsagon/create.py,sha256=BERrBviwMvifg5OwApqdanvULJHHk39fIvnTCZN3Xkk,4432
 parsagon/custom_function.py,sha256=oEj28qItaHUnsvLIHD7kg5QL3J3aO6rW6xKKP-H-Drs,770
 parsagon/edit.py,sha256=5gtnx0gNB7Gvz8ET00SczE-ZS0TomN1um6uObP-OObE,3120
 parsagon/exceptions.py,sha256=tG1vnpmUN1GdJ1GSpe1MaWH3zWmFLZCwtOfEGu8qPP0,910
-parsagon/executor.py,sha256=_jPdLxclmLekb03JCtxvslRpfUhVotL6rnHFw9EibXE,29927
+parsagon/executor.py,sha256=iKSUHJ4yu-6YHerenIJfFuhoX100OmdDGemkBJZ03q8,30260
 parsagon/gui_entry.py,sha256=bqG9K0CArXWWwDGoT8aV17YLNM8MfjSf6SJ_B3QbNeA,671
 parsagon/highlights.js,sha256=2UDfUApblU9xtGgTLCq4X7rHRV0wcqDSSFZPmJS6fJg,16643
 parsagon/main.py,sha256=_ww08Fpf1S8ePXSDg2GqtmnQvdfSgMK_AiNq5CuhWko,10001
@@ -25,8 +25,8 @@ parsagon/tests/test_invalid_args.py,sha256=TAFdHGy92lUxjljPrtODOuEGVss6rn-F5GvEK
 parsagon/tests/test_pipeline_operations.py,sha256=aEwZNtIwOl9X7jdLDLB4YEdgMp7_x8PXCINAE7RT4NY,805
 parsagon/tests/test_print.py,sha256=BG7f55YDBoL0S7k291-so_Gje_hUAQOkB-jh-bEYsJY,198
 parsagon/tests/test_secrets.py,sha256=Ctsscl2tmMTZcFAy5dnyqUlgTov2UharZgLpbRCLdEg,2662
-parsagon-0.17.1.dist-info/METADATA,sha256=-ylaryWc8rMx_nq9m3Oz_WUsQ7Ygw37pCQqnQn-1hkc,2567
-parsagon-0.17.1.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-parsagon-0.17.1.dist-info/entry_points.txt,sha256=I1UlPUb4oY2k9idkI8kvdkEcrjKGRSOl5pMbA6uu6kw,48
-parsagon-0.17.1.dist-info/top_level.txt,sha256=ih5uYQzW4qjhRKppys-WiHLIbXVZ99YdqDcfAtlcQwk,9
-parsagon-0.17.1.dist-info/RECORD,,
+parsagon-0.17.2.dist-info/METADATA,sha256=v_vLvOpQcU4xM9QAq_XeTRu_19ICXOVCoVSA8YCgrdM,2633
+parsagon-0.17.2.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+parsagon-0.17.2.dist-info/entry_points.txt,sha256=I1UlPUb4oY2k9idkI8kvdkEcrjKGRSOl5pMbA6uu6kw,48
+parsagon-0.17.2.dist-info/top_level.txt,sha256=ih5uYQzW4qjhRKppys-WiHLIbXVZ99YdqDcfAtlcQwk,9
+parsagon-0.17.2.dist-info/RECORD,,