fix: standardize model prefixes in rag_tool and update changelog

aritraroy24 · aritraroy24 · commit 2df5f0a2d8d7 · 2026-04-02T11:11:17.000+01:00
diff --git a/.gitignore b/.gitignore
@@ -186,10 +186,6 @@ applications
 vlm_test
 examples/vlm_piezo_test
 
-# Claude files
-CLAUDE.md
-.claude
-
 # Test results
 db
 results
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -24,16 +24,20 @@
 - `process_articles()` now routes user-provided `doi_list` by `general_publisher` from metadata and sends each DOI only to its matching source processor.
 
 ---
-## [0.1.6] - 02-04-2026
+## [0.1.6] - 2026-04-02
 ### Changed
 - Updated [README.md](README.md), [CITATION.cff](CITATION.cff) and docs with the published version (advance article) of the ComProScanner paper in _Digital Discovery_ as fully open access:
   - [ComProScanner: a multi-agent based framework for composition-property structured data extraction from scientific literature](https://doi.org/10.1039/D5DD00521C) 
 
 ### Added
 - Guide for API key creation for various LLM providers and publisher APIs added to the documentation at `docs/getting-started/api-key-guide.md` with detailed instructions for each provider.
 
+### Fixed
+- Model prefix handling in `rag_tool.py` standardized to reflect the docs.
+- `HF_TOKEN` documentation clarified as optional — only required for gated or private Hugging Face models.
+
 ---
-## [0.1.5] - 08-02-2026
+## [0.1.5] - 2026-02-08
 
 ### Added
 - Data related to comparison with other agentic data extraction frameworks added for the ComProScanner paper in the `examples/piezo_test/comparing_existing_frameworks` folder.
@@ -105,7 +109,7 @@
 - README badges section converted from HTML to markdown format for better compatibility across platforms.
 
 ---
-## [0.1.4] - 02-12-2025
+## [0.1.4] - 2025-12-02
 
 ### Added
 
@@ -141,7 +145,7 @@
   - [ComProScanner Workflow](https://raw.githubusercontent.com/aritraroy24/ComProScanner/main/assets/overall_workflow.png)
 
 ---
-## [0.1.3] - 04-11-2025
+## [0.1.3] - 2025-11-04
 
 ### Fixed
 
@@ -150,15 +154,15 @@
   - To `from langchain.text_splitter.recursive_character import RecursiveCharacterTextSplitter`
 
 ---
-## [0.1.2] - 24-10-2025
+## [0.1.2] - 2025-10-24
 
 ### Added
 
 - Link to ComProScanner preprint on arXiv in the documentation index page and README.md:
   - [arXiv:2510.20362](https://arxiv.org/abs/2510.20362)
 
 ---
-## [0.1.1] - 22-10-2025
+## [0.1.1] - 2025-10-22
 
 ### Fixed
 
@@ -167,7 +171,7 @@
   - [ComProScanner Workflow](https://i.ibb.co/QWd2qd3/overall-workflow.png)
 
 ---
-## [0.1.0] - 22-10-2025
+## [0.1.0] - 2025-10-22
 
 ### Added
 
diff --git a/docs/about/changelog.md b/docs/about/changelog.md
@@ -24,7 +24,7 @@
 - `process_articles()` now routes user-provided `doi_list` by `general_publisher` from metadata and sends each DOI only to its matching source processor.
 
 ---
-## [0.1.6] - 02-04-2026
+## [0.1.6] - 2026-04-02
 ### Changed
 - Updated [README.md](README.md), [CITATION.cff](CITATION.cff) and docs with the published version (advance article) of the ComProScanner paper in _Digital Discovery_ as fully open access:
   - [ComProScanner: a multi-agent based framework for composition-property structured data extraction from scientific literature](https://doi.org/10.1039/D5DD00521C) 
@@ -33,7 +33,7 @@
 - Guide for API key creation for various LLM providers and publisher APIs added to the documentation at `docs/getting-started/api-key-guide.md` with detailed instructions for each provider.
 
 ---
-## [0.1.5] - 08-02-2026
+## [0.1.5] - 2026-02-08
 
 ### Added
 - Data related to comparison with other agentic data extraction frameworks added for the ComProScanner paper in the `examples/piezo_test/comparing_existing_frameworks` folder.
@@ -105,7 +105,7 @@
 - README badges section converted from HTML to markdown format for better compatibility across platforms.
 
 ---
-## [0.1.4] - 02-12-2025
+## [0.1.4] - 2025-12-02
 
 ### Added
 
@@ -141,7 +141,7 @@
   - [ComProScanner Workflow](https://raw.githubusercontent.com/aritraroy24/ComProScanner/main/assets/overall_workflow.png)
 
 ---
-## [0.1.3] - 04-11-2025
+## [0.1.3] - 2025-11-04
 
 ### Fixed
 
@@ -150,15 +150,15 @@
   - To `from langchain.text_splitter.recursive_character import RecursiveCharacterTextSplitter`
 
 ---
-## [0.1.2] - 24-10-2025
+## [0.1.2] - 2025-10-24
 
 ### Added
 
 - Link to ComProScanner preprint on arXiv in the documentation index page and README.md:
   - [arXiv:2510.20362](https://arxiv.org/abs/2510.20362)
 
 ---
-## [0.1.1] - 22-10-2025
+## [0.1.1] - 2025-10-22
 
 ### Fixed
 
@@ -167,7 +167,7 @@
   - [ComProScanner Workflow](https://i.ibb.co/QWd2qd3/overall-workflow.png)
 
 ---
-## [0.1.0] - 22-10-2025
+## [0.1.0] - 2025-10-22
 
 ### Added
 
diff --git a/docs/getting-started/api-key-guide.md b/docs/getting-started/api-key-guide.md
@@ -200,7 +200,7 @@ OPENROUTER_API_KEY=your_openrouter_api_key
 
 Environment variable: `TOGETHER_API_KEY`
 
-Typical model prefixes: `together_ai/...`
+Typical model prefixes: `together/...`
 
 How to get it:
 
@@ -232,7 +232,7 @@ COHERE_API_KEY=your_cohere_api_key
 
 Environment variable: `FIREWORKS_API_KEY`
 
-Typical model prefixes: `fireworks_ai/...`
+Typical model prefixes: `fireworks/...`
 
 How to get it:
 
@@ -264,11 +264,12 @@ How to set it up:
 
 Environment variable: `HF_TOKEN`
 
+> **Optional.** Only required for downloading gated or private Hugging Face models. Public models work without a token.
+
 Used for:
 
-- Accessing the default Hugging Face embedding model workflow
-- Accessing gated or rate-limited Hugging Face models
-- Optional embedding/model downloads when required
+- Accessing gated or private Hugging Face models
+- Rate-limited API access
 
 How to get it:
 
diff --git a/docs/rag-config.md b/docs/rag-config.md
@@ -123,7 +123,7 @@ scanner.extract_composition_property_data(
 scanner.extract_composition_property_data(
     main_extraction_keyword="d33",
     rag_db_path="embeddings/piezo",
-    rag_chat_model="deepseek-chat",
+    rag_chat_model="deepseek/deepseek-chat",
     rag_max_tokens=1024,
     rag_top_k=4,
 )
@@ -178,7 +178,7 @@ scanner.extract_composition_property_data(
 scanner.extract_composition_property_data(
     main_extraction_keyword="d33",
     rag_db_path="embeddings/piezo",
-    rag_chat_model="together_ai/meta-llama/Llama-3-70b-chat-hf",
+    rag_chat_model="together/meta-llama/Llama-3-70b-chat-hf",
     rag_max_tokens=1024,
     rag_top_k=4,
 )
@@ -220,7 +220,7 @@ scanner.extract_composition_property_data(
 scanner.extract_composition_property_data(
     main_extraction_keyword="d33",
     rag_db_path="embeddings/piezo",
-    rag_chat_model="fireworks_ai/accounts/fireworks/models/llama-v3-8b-instruct",
+    rag_chat_model="fireworks/models/llama-v3-8b-instruct",
     rag_max_tokens=1024,
     rag_top_k=4,
 )
diff --git a/src/comproscanner/extract_flow/tools/rag_tool.py b/src/comproscanner/extract_flow/tools/rag_tool.py
@@ -82,28 +82,28 @@ def _get_llm(self) -> BaseChatModel:
             "callbacks": callbacks,
         }
         # OpenAI models
-        if model.startswith(("gpt-", "text-", "o1", "o3")):
+        if model.startswith(("openai/", "gpt-", "text-", "o1", "o3")):
             self._check_package_exists("langchain_openai", model)
             from langchain_openai import ChatOpenAI
 
             return ChatOpenAI(model=model, request_timeout=1000, **common_params)
 
         # Deepseek models
-        if model.startswith("deepseek"):
+        if model.startswith("deepseek/"):
             self._check_package_exists("langchain_deepseek", model)
             from langchain_deepseek import ChatDeepSeek
 
             return ChatDeepSeek(model=model, request_timeout=1000, **common_params)
 
         # Google Gemini models
-        elif model.startswith("gemini-"):
+        elif model.startswith("gemini/"):
             self._check_package_exists("langchain_google_genai", model)
             from langchain_google_genai import ChatGoogleGenerativeAI
 
             return ChatGoogleGenerativeAI(model=model, **common_params)
 
         # Anthropic Claude models
-        elif model.startswith("claude-"):
+        elif model.startswith("claude/"):
             self._check_package_exists("langchain_anthropic", model)
             from langchain_anthropic import ChatAnthropic
 
@@ -143,7 +143,7 @@ def _get_llm(self) -> BaseChatModel:
             return ChatCohere(model=model_name, **common_params)
 
         # Fireworks models
-        elif model.startswith(("fireworks/", "accounts/fireworks")):
+        elif model.startswith(("fireworks/")):
             self._check_package_exists("langchain_fireworks", model)
             from langchain_fireworks import ChatFireworks
 
