LLM attribute calculation (#70)

* perf: add LLM_RESPONSE support

* fix: overwriting default user prompt with ac.USER_PROMPT

* perf: mustache align to requirements

* perf: avoid variable overrides by end users

* fix: mustache rendering

* feat: implement asynchronous processing for LLM_RESPONSE data type

* fix: update batch size calculation to use NUM_WORKERS_A2VYBG

* ensure at least 1 worker exists

* fix: update progress tracking in LLM record processing

* fix: improve progress tracking in LLM record processing

* perf: unprototype run_ac.py

* perf: add llm ac caching

* adds valid cache link check which fixes playground error

* fix: clarify log message for cached record usage

* refactor: streamline attribute calculation and caching logic

* refactor: rename functions and variables for clarity and consistency + cache at end

* chore: pr review

* chore: pr review

---------

Co-authored-by: Moritz Feuerpfeil <[email protected]>
Co-authored-by: JWittmeyer <[email protected]>

Author: 3 people

mustache.py

@@ -0,0 +1,9 @@
+from typing import Dict, Any
+import chevron
+
+
+def prepare_and_render_mustache(template_text: str, record_dict: Dict[str, Any]) -> str:
+    return chevron.render(
+        template=template_text,
+        data=record_dict["data"],
+    )

requirements.txt

@@ -9,9 +9,7 @@ aiohappyeyeballs==2.4.3
     #   -r requirements/exec-env-requirements.txt
     #   aiohttp
 aiohttp==3.11.7
-    # via
-    #   -r requirements/exec-env-requirements.txt
-    #   openai
+    # via -r requirements/exec-env-requirements.txt
 aiosignal==1.3.1
     # via
     #   -r requirements/exec-env-requirements.txt
@@ -23,6 +21,7 @@ anyio==4.6.2.post1
     #   -r requirements/exec-env-requirements.txt
     #   groq
     #   httpx
+    #   openai
 async-timeout==5.0.1
     # via
     #   -r requirements/exec-env-requirements.txt
@@ -68,6 +67,8 @@ charset-normalizer==3.4.0
     # via
     #   -r requirements/exec-env-requirements.txt
     #   requests
+chevron==0.14.0
+    # via -r requirements/requirements.in
 click==8.1.7
     # via
     #   -r requirements/exec-env-requirements.txt
@@ -99,6 +100,7 @@ distro==1.9.0
     # via
     #   -r requirements/exec-env-requirements.txt
     #   groq
+    #   openai
 docopt==0.6.2
     # via
     #   -r requirements/exec-env-requirements.txt
@@ -144,6 +146,7 @@ httpx==0.27.2
     # via
     #   -r requirements/exec-env-requirements.txt
     #   groq
+    #   openai
 idna==3.10
     # via
     #   -r requirements/exec-env-requirements.txt
@@ -167,6 +170,8 @@ jinja2==3.1.4
     # via
     #   -r requirements/exec-env-requirements.txt
     #   spacy
+jiter==0.8.2
+    # via openai
 joblib==1.4.2
     # via
     #   -r requirements/exec-env-requirements.txt
@@ -254,8 +259,8 @@ numpy==1.23.4
     #   spacy
     #   textacy
     #   thinc
-openai==0.27.7
-    # via -r requirements/exec-env-requirements.txt
+openai==1.59.7
+    # via -r requirements/requirements.in
 packaging==24.2
     # via
     #   -r requirements/exec-env-requirements.txt
@@ -297,6 +302,7 @@ pydantic==2.7.4
     #   -r requirements/requirements.in
     #   confection
     #   groq
+    #   openai
     #   spacy
     #   thinc
     #   weasel
@@ -343,7 +349,6 @@ requests==2.31.0
     #   -r requirements/exec-env-requirements.txt
     #   google-search-results
     #   lexmo
-    #   openai
     #   spacy
     #   sumy
     #   textacy
@@ -380,6 +385,7 @@ sniffio==1.3.1
     #   anyio
     #   groq
     #   httpx
+    #   openai
 soupsieve==2.6
     # via
     #   -r requirements/exec-env-requirements.txt
@@ -466,6 +472,7 @@ typing-extensions==4.12.2
     #   cloudpathlib
     #   groq
     #   multidict
+    #   openai
     #   pydantic
     #   pydantic-core
     #   sqlalchemy

requirements/exec-env-requirements.txt

@@ -171,7 +171,7 @@ numpy==1.23.4
     #   spacy
     #   textacy
     #   thinc
-openai==0.27.7
+# openai==0.27.7
     # via -r exec-env-requirements.in
 packaging==24.2
     # via

requirements/requirements.in

@@ -1,3 +1,5 @@
 -r exec-env-requirements.txt
 pydantic==2.7.4
-spacy[ja]==3.7.5
+spacy[ja]==3.7.5
+chevron==0.14.0
+openai==1.59.7