Merge remote-tracking branch 'upstream/main' into linting

.env.example

@@ -65,4 +65,12 @@ LMSTUDIO_API_BASE_URL=
 XAI_API_KEY=
 
 # Include this environment variable if you want more logging for debugging locally
-VITE_LOG_LEVEL=debug
+VITE_LOG_LEVEL=debug
+
+# Example Context Values for qwen2.5-coder:32b
+# 
+# DEFAULT_NUM_CTX=32768 # Consumes 36GB of VRAM
+# DEFAULT_NUM_CTX=24576 # Consumes 32GB of VRAM
+# DEFAULT_NUM_CTX=12288 # Consumes 26GB of VRAM
+# DEFAULT_NUM_CTX=6144 # Consumes 24GB of VRAM
+DEFAULT_NUM_CTX=

CONTRIBUTING.md

@@ -1,4 +1,7 @@
 # Contributing to Bolt.new Fork
+## DEFAULT_NUM_CTX
+
+The `DEFAULT_NUM_CTX` environment variable can be used to limit the maximum number of context values used by the qwen2.5-coder model. For example, to limit the context to 24576 values (which uses 32GB of VRAM), set `DEFAULT_NUM_CTX=24576` in your `.env.local` file.
 
 First off, thank you for considering contributing to Bolt.new! This fork aims to expand the capabilities of the original project by integrating multiple LLM providers and enhancing functionality. Every contribution helps make Bolt.new a better tool for developers worldwide.
 
@@ -81,6 +84,19 @@ ANTHROPIC_API_KEY=XXX
 ```bash
 VITE_LOG_LEVEL=debug
 ```
+
+   - Optionally set context size:
+```bash
+DEFAULT_NUM_CTX=32768
+```
+
+Some Example Context Values for the qwen2.5-coder:32b models are.
+ 
+* DEFAULT_NUM_CTX=32768 - Consumes 36GB of VRAM
+* DEFAULT_NUM_CTX=24576 - Consumes 32GB of VRAM
+* DEFAULT_NUM_CTX=12288 - Consumes 26GB of VRAM
+* DEFAULT_NUM_CTX=6144 - Consumes 24GB of VRAM
+
 **Important**: Never commit your `.env.local` file to version control. It's already included in .gitignore.
 
 ### 🚀 Running the Development Server

Dockerfile

@@ -26,6 +26,7 @@ ARG OPEN_ROUTER_API_KEY
 ARG GOOGLE_GENERATIVE_AI_API_KEY
 ARG OLLAMA_API_BASE_URL
 ARG VITE_LOG_LEVEL=debug
+ARG DEFAULT_NUM_CTX
 
 ENV WRANGLER_SEND_METRICS=false \
     GROQ_API_KEY=${GROQ_API_KEY} \
@@ -35,7 +36,8 @@ ENV WRANGLER_SEND_METRICS=false \
     OPEN_ROUTER_API_KEY=${OPEN_ROUTER_API_KEY} \
     GOOGLE_GENERATIVE_AI_API_KEY=${GOOGLE_GENERATIVE_AI_API_KEY} \
     OLLAMA_API_BASE_URL=${OLLAMA_API_BASE_URL} \
-    VITE_LOG_LEVEL=${VITE_LOG_LEVEL}
+    VITE_LOG_LEVEL=${VITE_LOG_LEVEL} \
+    DEFAULT_NUM_CTX=${DEFAULT_NUM_CTX}
 
 # Pre-configure wrangler to disable metrics
 RUN mkdir -p /root/.config/.wrangler && \
@@ -57,6 +59,7 @@ ARG OPEN_ROUTER_API_KEY
 ARG GOOGLE_GENERATIVE_AI_API_KEY
 ARG OLLAMA_API_BASE_URL
 ARG VITE_LOG_LEVEL=debug
+ARG DEFAULT_NUM_CTX
 
 ENV GROQ_API_KEY=${GROQ_API_KEY} \
     HuggingFace_API_KEY=${HuggingFace_API_KEY} \
@@ -65,7 +68,8 @@ ENV GROQ_API_KEY=${GROQ_API_KEY} \
     OPEN_ROUTER_API_KEY=${OPEN_ROUTER_API_KEY} \
     GOOGLE_GENERATIVE_AI_API_KEY=${GOOGLE_GENERATIVE_AI_API_KEY} \
     OLLAMA_API_BASE_URL=${OLLAMA_API_BASE_URL} \
-    VITE_LOG_LEVEL=${VITE_LOG_LEVEL}
+    VITE_LOG_LEVEL=${VITE_LOG_LEVEL} \
+    DEFAULT_NUM_CTX=${DEFAULT_NUM_CTX}
 
 RUN mkdir -p ${WORKDIR}/run
 CMD pnpm run dev --host

app/lib/.server/llm/model.ts

@@ -11,7 +11,11 @@ import { createOpenRouter } from '@openrouter/ai-sdk-provider';
 import { createMistral } from '@ai-sdk/mistral';
 import { createCohere } from '@ai-sdk/cohere';
 
-export function getAnthropicModel(apiKey: string | undefined, model: string) {
+export const DEFAULT_NUM_CTX = process.env.DEFAULT_NUM_CTX ?
+  parseInt(process.env.DEFAULT_NUM_CTX, 10) :
+  32768;
+
+export function getAnthropicModel(apiKey: string, model: string) {
   const anthropic = createAnthropic({
     apiKey,
   });
@@ -68,7 +72,7 @@ export function getGroqModel(apiKey: string | undefined, model: string) {
   return openai(model);
 }
 
-export function getHuggingFaceModel(apiKey: string | undefined, model: string) {
+export function getHuggingFaceModel(apiKey: string, model: string) {
   const openai = createOpenAI({
     baseURL: 'https://api-inference.huggingface.co/v1/',
     apiKey,
@@ -78,8 +82,8 @@ export function getHuggingFaceModel(apiKey: string | undefined, model: string) {
 }
 
 export function getOllamaModel(baseURL: string, model: string) {
-  const ollamaInstance = ollama(model, {
-    numCtx: 32768,
+  let Ollama = ollama(model, {
+    numCtx: DEFAULT_NUM_CTX,
   });
 
   ollamaInstance.config.baseURL = `${baseURL}/api`;

docker-compose.yaml

@@ -21,6 +21,7 @@ services:
       - GOOGLE_GENERATIVE_AI_API_KEY=${GOOGLE_GENERATIVE_AI_API_KEY}
       - OLLAMA_API_BASE_URL=${OLLAMA_API_BASE_URL}
       - VITE_LOG_LEVEL=${VITE_LOG_LEVEL:-debug}
+      - DEFAULT_NUM_CTX=${DEFAULT_NUM_CTX:-32768}
       - RUNNING_IN_DOCKER=true
     extra_hosts:
       - "host.docker.internal:host-gateway"      
@@ -48,6 +49,7 @@ services:
       - GOOGLE_GENERATIVE_AI_API_KEY=${GOOGLE_GENERATIVE_AI_API_KEY}
       - OLLAMA_API_BASE_URL=${OLLAMA_API_BASE_URL}
       - VITE_LOG_LEVEL=${VITE_LOG_LEVEL:-debug}
+      - DEFAULT_NUM_CTX=${DEFAULT_NUM_CTX:-32768}
       - RUNNING_IN_DOCKER=true
     extra_hosts:
       - "host.docker.internal:host-gateway"