Fix #948 - don't try to install spaCy at runtime or require it

Author: nonprofittechy

docassemble/ALWeaver/data/questions/assembly_line.yml

@@ -59,10 +59,6 @@ code: |
         if yes_recognize_form_fields: # ask to add fields
           process_field_recognition
       if im_feeling_lucky:
-        if not formfyxer_available() and user_logged_in() and user_has_privilege(["admin"]) and install_en_core_web_lg:
-          install_spacy_background_task
-          if not install_spacy_background_task.ready():
-            install_spacy_background_waiting_screen
         nav.set_sections([
           {"nav_generate_interview": "Generate your interview"},
           {"nav_review": "Review your answers"},
@@ -1904,20 +1900,6 @@ subquestion: |
 
   % endfor
 
-  % if not formfyxer_available():
-  <div class="alert alert-warning" role="alert">
-  Automatic field grouping isn't set up. To use it, you need:
-
-  <ol>
-  <li>The `en_core_web_lg` installed, or</li>
-  <li>An API token for tools.suffolklitlab.org installed in your configuration.</li>
-  </ol>
-
-  You can still use auto drafting mode but all questions will be
-  placed on one screen. You can edit them in your playground later.
-  </div>
-  % endif
-
 fields: 
   - Do you want us to draft the interview for you?: im_feeling_lucky
     datatype: radio
@@ -1947,40 +1929,6 @@ fields:
     show if: interview.start_with_json
     accept: |
       "application/json"
-  - Try to install the `en_core_web_lg` package before using auto drafting mode: install_en_core_web_lg
-    help: |
-      Installing `en_core_web_lg` will allow you to use automatic field
-      grouping but it may use more memory on your server.
-
-      If you are a nonprofit, you can request an API token to use
-      auto field grouping without installing `en_core_web_lg`.
-    datatype: yesnoradio
-    show if:
-      variable: im_feeling_lucky
-      is: True
-      code: |
-        not formfyxer_available() and user_logged_in() and user_has_privilege(["admin"])
----
-code: |
-  install_spacy_background_task = background_action("install_spacy_background_action")
----
-code: |
-  install_spacy_model()
-  background_response()
-  install_spacy_background_action = True
----
-event: install_spacy_background_waiting_screen
-id: waiting screen
-question: |
-  Installing `en_core_web_lg`
-subquestion: |
-  Please wait while we install `en_core_web_lg`. This can take a
-  few minutes to complete.
-
-  <div class="spinner-border" role="status">
-    <span class="visually-hidden">Loading...</span>
-  </div>
-reload: True
 ---
 template: explain_auto_drafting
 subject: |

docassemble/ALWeaver/interview_generator.py

@@ -58,19 +58,12 @@
 import re
 import uuid
 import zipfile
-import spacy
 from dataclasses import dataclass
 import pycountry
 
 mako.runtime.UNDEFINED = DAEmpty()
 
 
-def formfyxer_available():
-    if get_config("assembly line", {}).get("tools.suffolklitlab.org api key"):
-        return True
-    return spacy.util.is_package("en_core_web_lg")
-
-
 TypeType = type(type(None))
 
 __all__ = [
@@ -87,7 +80,6 @@ def formfyxer_available():
     "escape_quotes",
     "field_type_options",
     "fix_id",
-    "formfyxer_available",
     "get_character_limit",
     "get_court_choices",
     "get_docx_validation_errors",
@@ -98,7 +90,6 @@ def formfyxer_available():
     "get_pdf_variable_name_matches",
     "get_variable_name_warnings",
     "indent_by",
-    "install_spacy_model",
     "is_reserved_docx_label",
     "is_reserved_label",
     "is_url",
@@ -146,11 +137,6 @@ def formfyxer_available():
 remove_u = re.compile(r"^u")
 
 
-def install_spacy_model(model="en_core_web_lg"):
-    if not spacy.util.is_package(model):
-        spacy.cli.download(model)
-
-
 class ParsingException(Exception):
     """Throws an error if we can't understand the labels somehow, so we can tell the user"""
 
@@ -1803,23 +1789,22 @@ def _guess_categories(self, title) -> List[str]:
         Returns:
             List[str]: A list of categories
         """
-        if formfyxer_available():
-            # Get the full text of all templates
-            full_text = ""
-            for template in self.uploaded_templates:
-                if template.filename.lower().endswith(".pdf"):
-                    full_text += extract_text(template.path())
-                elif template.filename.lower().endswith(".docx"):
-                    docx_data = docx2python(
-                        template.path()
-                    )  # Will error with invalid value
-                    full_text += docx_data.text
-            categories = formfyxer.spot(
-                title + ": " + full_text,
-                token=get_config("assembly line", {}).get("spot api key", None),
-            )
-            if categories and not "401" in categories:
-                return categories
+        # Get the full text of all templates
+        full_text = ""
+        for template in self.uploaded_templates:
+            if template.filename.lower().endswith(".pdf"):
+                full_text += extract_text(template.path())
+            elif template.filename.lower().endswith(".docx"):
+                docx_data = docx2python(
+                    template.path()
+                )  # Will error with invalid value
+                full_text += docx_data.text
+        categories = formfyxer.spot(
+            title + ": " + full_text,
+            token=get_config("assembly line", {}).get("spot api key", None),
+        )
+        if categories and not "401" in categories:
+            return categories
         # Top hits: Housing, Family, Consumer, Probate, Criminal, Traffic, Consumer, Health, Immigration, Employment
         if any(
             keyword in title.lower()

setup.py

@@ -53,7 +53,7 @@ def find_package_data(where='.', package='', exclude=standard_exclude, exclude_d
       url='https://docassemble.org',
       packages=find_packages(),
       namespace_packages=['docassemble'],
-      install_requires=['PyYAML>=5.1.2', 'beautifulsoup4>=4.11.1', 'docassemble.ALToolbox>=0.4.2', 'docassemble.AssemblyLine>=2.11.3', 'docx2python>=1.27.1', 'more-itertools>=8.6.0', 'numpy>=1.0.4', 'pikepdf>=5.1.1', 'scikit-learn>=0.0', 'spacy>=3.2.0', 'formfyxer>=0.3.0a1'],
+      install_requires=['PyYAML>=5.1.2', 'beautifulsoup4>=4.11.1', 'docassemble.ALToolbox>=0.4.2', 'docassemble.AssemblyLine>=2.11.3', 'docx2python>=1.27.1', 'more-itertools>=8.6.0', 'numpy>=1.0.4', 'pikepdf>=5.1.1', 'formfyxer>=0.3.0a1'],
       zip_safe=False,
       package_data=find_package_data(where='docassemble/ALWeaver/', package='docassemble.ALWeaver'),
      )