Default to gemini 2.0 flash instead of 1.5 flash on Gemini setup

Add price of gemini 2.0 flash for cost calculations

Author: debanjum

.github/workflows/run_evals.yml

@@ -147,7 +147,7 @@ jobs:
           echo "## Evaluation Summary of Khoj on ${{ matrix.dataset }} in ${{ matrix.khoj_mode }} mode" >> $GITHUB_STEP_SUMMARY
           echo "**$(head -n 1 *_evaluation_summary_*.txt)**" >> $GITHUB_STEP_SUMMARY
           echo "- Khoj Version: ${{ steps.hatch.outputs.version }}" >> $GITHUB_STEP_SUMMARY
-          echo "- Chat Model: Gemini 1.5 Flash 002" >> $GITHUB_STEP_SUMMARY
+          echo "- Chat Model: Gemini 2.0 Flash" >> $GITHUB_STEP_SUMMARY
           echo "\`\`\`" >> $GITHUB_STEP_SUMMARY
           tail -n +2 *_evaluation_summary_*.txt >> $GITHUB_STEP_SUMMARY
           echo "" >> $GITHUB_STEP_SUMMARY

documentation/docs/get-started/setup.mdx

@@ -333,7 +333,7 @@ Using Ollama? See the [Ollama Integration](/advanced/ollama) section for more cu
    - Add your [Gemini API key](https://aistudio.google.com/app/apikey)
    - Give the configuration a friendly name like `Gemini`. Do not configure the API base url.
 2. Create a new [chat model](http://localhost:42110/server/admin/database/chatmodel/add)
-    - Set the `chat-model` field to a [Google Gemini chat model](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models#gemini-models). Example: `gemini-1.5-flash`.
+    - Set the `chat-model` field to a [Google Gemini chat model](https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models#gemini-models). Example: `gemini-2.0-flash`.
     - Set the `model-type` field to `Google`.
     - Set the `ai model api` field to the Gemini AI Model API you created in step 1.
 

src/khoj/processor/conversation/google/gemini_chat.py

@@ -31,7 +31,7 @@
 
 def extract_questions_gemini(
     text,
-    model: Optional[str] = "gemini-1.5-flash",
+    model: Optional[str] = "gemini-2.0-flash",
     conversation_log={},
     api_key=None,
     temperature=0,
@@ -132,9 +132,9 @@ def gemini_send_message_to_model(
 
     model_kwargs = {}
 
-    # Sometimes, this causes unwanted behavior and terminates response early. Disable for now while it's flaky.
-    # if response_type == "json_object":
-    #     model_kwargs["response_mime_type"] = "application/json"
+    # This caused unwanted behavior and terminates response early for gemini 1.5 series. Monitor for flakiness with 2.0 series.
+    if response_type == "json_object" and model in ["gemini-2.0-flash"]:
+        model_kwargs["response_mime_type"] = "application/json"
 
     # Get Response from Gemini
     return gemini_completion_with_backoff(
@@ -154,7 +154,7 @@ def converse_gemini(
     online_results: Optional[Dict[str, Dict]] = None,
     code_results: Optional[Dict[str, Dict]] = None,
     conversation_log={},
-    model: Optional[str] = "gemini-1.5-flash",
+    model: Optional[str] = "gemini-2.0-flash",
     api_key: Optional[str] = None,
     temperature: float = 0.2,
     completion_func=None,

src/khoj/utils/constants.py

@@ -18,7 +18,7 @@
     "bartowski/Qwen2.5-14B-Instruct-GGUF",
 ]
 default_openai_chat_models = ["gpt-4o-mini", "gpt-4o"]
-default_gemini_chat_models = ["gemini-1.5-flash", "gemini-1.5-pro"]
+default_gemini_chat_models = ["gemini-2.0-flash", "gemini-1.5-pro"]
 default_anthropic_chat_models = ["claude-3-5-sonnet-20241022", "claude-3-5-haiku-20241022"]
 
 empty_config = {
@@ -46,6 +46,7 @@
     "gemini-1.5-flash-002": {"input": 0.075, "output": 0.30},
     "gemini-1.5-pro": {"input": 1.25, "output": 5.00},
     "gemini-1.5-pro-002": {"input": 1.25, "output": 5.00},
+    "gemini-2.0-flash": {"input": 0.10, "output": 0.40},
     # Anthropic Pricing: https://www.anthropic.com/pricing#anthropic-api_
     "claude-3-5-sonnet-20241022": {"input": 3.0, "output": 15.0},
     "claude-3-5-haiku-20241022": {"input": 1.0, "output": 5.0},

tests/conftest.py

@@ -315,7 +315,7 @@ def chat_client_builder(search_config, user, index_content=True, require_auth=Fa
     if chat_provider == ChatModel.ModelType.OPENAI:
         online_chat_model = ChatModelFactory(name="gpt-4o-mini", model_type="openai")
     elif chat_provider == ChatModel.ModelType.GOOGLE:
-        online_chat_model = ChatModelFactory(name="gemini-1.5-flash", model_type="google")
+        online_chat_model = ChatModelFactory(name="gemini-2.0-flash", model_type="google")
     elif chat_provider == ChatModel.ModelType.ANTHROPIC:
         online_chat_model = ChatModelFactory(name="claude-3-5-haiku-20241022", model_type="anthropic")
     if online_chat_model:

tests/evals/eval.py

@@ -629,7 +629,7 @@ def main():
         response_evaluator = evaluate_response_with_mcq_match
     elif args.dataset == "math500":
         response_evaluator = partial(
-            evaluate_response_with_gemini, eval_model=os.getenv("GEMINI_EVAL_MODEL", "gemini-1.5-flash-002")
+            evaluate_response_with_gemini, eval_model=os.getenv("GEMINI_EVAL_MODEL", "gemini-2.0-flash-001")
         )
     elif args.dataset == "frames_ir":
         response_evaluator = evaluate_response_for_ir