Merge pull request #1822 from laws-africa/scripts

this includes pdfjs-to-text in .env/bin so that it's available as an …

peachjam/helpers.py

@@ -1,11 +1,11 @@
-import os
 import string
 import subprocess
 import tempfile
 from datetime import date, datetime
 from functools import wraps
 
 import martor.utils
+from django.conf import settings
 from django.utils.translation import get_language_from_request
 from languages_plus.models import Language
 
@@ -46,7 +46,7 @@ def pdfjs_to_text(fname):
     """Extract text from fname using pdfjs-compatible script."""
     with tempfile.NamedTemporaryFile(suffix=".txt") as outf:
         cmd = [
-            os.path.join(os.path.dirname(__file__), "..", "bin", "pdfjs-to-text"),
+            settings.PEACHJAM["PDFJS_TO_TEXT"],
             fname,
             outf.name,
         ]

peachjam/settings.py

@@ -130,6 +130,7 @@
     "SEARCH_JURISDICTION_FILTER": False,
     "MULTIPLE_JURISDICTIONS": False,
     "MULTIPLE_LOCALITIES": False,
+    "PDFJS_TO_TEXT": "bin/pdfjs-to-text",
 }
 
 PEACHJAM["ES_INDEX"] = os.environ.get("ES_INDEX", slugify(PEACHJAM["APP_NAME"]))

pyproject.toml

@@ -70,3 +70,6 @@ dependencies = [
     "sentry-sdk>=1.16.0",
     "whitenoise>=6.0.0",
 ]
+
+[tool.setuptools.data-files]
+"bin" = ["bin/pdfjs-to-text"]