Feature/add rag to synthesizer (#133)

* add output * remove broken pdf build * update model * tally distribution * add rag
SciPhi-AI · Dec 19, 2023 · fffcb67 · fffcb67
1 parent fdca94b
commit fffcb67
Show file tree

Hide file tree

Showing 7 changed files with 570 additions and 12 deletions.
diff --git a/docs/source/setup/quickstart.rst b/docs/source/setup/quickstart.rst
@@ -26,8 +26,30 @@ For additional details, refer to the `installation guide <https://sciphi.readthe
 
 Using Synthesizer
 -----------------
+1. **Run with RAG**
 
-1. **Generate synthetic question answer pairs**
+   .. code-block:: shell
+
+      export SCIPHI_API_KEY=MY_SCIPHI_API_KEY
+      # Queries AgentSearch and passes results on to the SciPhi Sensei
+      python -m agent_search.scripts.run_rag run --query="Your Search Query"  --rag_provider_name="agent-search" 
+      
+   .. code-block:: shell
+
+      export SCIPHI_API_KEY=MY_SCIPHI_API_KEY
+      export SERP_API_KEY=MY_SERP_API_KEY # refer to https://serpapi.com
+      # Queries Google and passes results on to the SciPhi Sensei
+      python -m agent_search.scripts.run_rag run --query="Your Search Query" --rag_provider_name=google-search
+      
+
+   .. code-block:: bash
+
+      tail augmented_output/config_name_eq_answer_question__dataset_name_eq_wiki_qa.jsonl
+      { "formatted_prompt": "... ### Question:\nwhat country did wine originate in\n\n### Input:\n1. URL: https://en.wikipedia.org/wiki/History%20of%20wine (Score: 0.85)\nTitle:History of wine....",
+      { "completion": Wine originated in the South Caucasus, which is now part of modern-day Armenia ...
+
+
+2. **Generate synthetic question answer pairs**
 
    .. code-block:: bash
 
@@ -40,7 +62,7 @@ Using Synthesizer
       { "formatted_prompt": "... ### Question:\nwhat country did wine originate in\n\n### Input:\n1. URL: https://en.wikipedia.org/wiki/History%20of%20wine (Score: 0.85)\nTitle:History of wine....",
       { "completion": Wine originated in the South Caucasus, which is now part of modern-day Armenia ...
 
-2. **Evaluate RAG pipeline performance**
+3. **Evaluate RAG pipeline performance**
 
    .. code-block:: bash
 

diff --git a/synthesizer/config/prompts/answer_question.yaml b/synthesizer/config/prompts/answer_question.yaml
diff --git a/synthesizer/core/base.py b/synthesizer/core/base.py
@@ -18,3 +18,4 @@ class LLMProviderName(Enum):
 class RAGProviderName(Enum):
     LOCAL = "local"
     AGENT_SEARCH = "agent-search"
+    GOOGLE_SEARCH = "google-search"
diff --git a/synthesizer/interface/__init__.py b/synthesizer/interface/__init__.py
@@ -16,6 +16,10 @@
     AgentSearchRAGConfig,
     AgentSearchRAGInterface,
 )
+from synthesizer.interface.rag.google_search import (
+    GoogleSearchRAGConfig,
+    GoogleSearchRAGInterface,
+)
 from synthesizer.interface.rag.local import LocalRAGInterface
 from synthesizer.interface.rag_interface_manager import RAGInterfaceManager
 
@@ -38,4 +42,6 @@
     "LocalRAGInterface",
     "AgentSearchRAGConfig",
     "AgentSearchRAGInterface",
+    "GoogleSearchRAGConfig",
+    "GoogleSearchRAGInterface",
 ]
diff --git a/synthesizer/interface/rag/agent_search.py b/synthesizer/interface/rag/agent_search.py
@@ -61,12 +61,3 @@ def get_rag_context(self, query) -> list[str]:
                 for i, result in enumerate(results)
             ]
         )
-
-        # logging.info(
-        #     f"{i}. \033[94mURL: {result.url}\033[0m (Score: \033[95m{result.score:.2f}\033[0m)"
-        # )
-        # logging.info("-" * 50)
-        # logging.info(f"Title: \033[93m{result.title}\033[0m")
-        # logging.info(f"Text:\n{result.text}\n")
-        # # logging.info(f"Metadata:\n{result.metadata}...")
-        # logging.info("-" * 80)