update readme and config

Author: carsontung666

README.md

@@ -43,7 +43,7 @@ import contextdb
 db = contextdb.open("my_docs.sqlite")
 
 # Configure LLM
-db.set_llm(provider="anthropic", model="claude-sonnet-4-20250514")
+db.set_llm(provider="anthropic", model="claude-sonnet-4-6")
 
 # Store a document tree
 tree_id = db.store(document_tree_json, format="document")
@@ -75,20 +75,31 @@ ct.close()
 
 ## Configuration
 
-Create a `.env` file:
+Create a `.env` file with your API keys:
 
 ```
 ANTHROPIC_API_KEY=sk-...
 OPENAI_API_KEY=sk-...
-LLM_PROVIDER=anthropic
-LLM_MODEL=claude-sonnet-4-20250514
 ```
 
-Or configure programmatically:
+Model and provider settings live in `contextdb/config/config.yaml`:
 
-```python
-from contextdb.config import Config
-llm = Config.get_llm_client()
+```yaml
+llm:
+  provider: anthropic          # anthropic or openai
+  model: claude-sonnet-4-6     # any model the provider supports
+  context_limit: 100000
+  max_concurrent: 10
+
+retriever:
+  beam_size: 3
+  max_turns: 5
+```
+
+Override at runtime with environment variables:
+
+```bash
+LLM_MODEL=claude-opus-4-6 python your_script.py
 ```
 
 ---
@@ -114,17 +125,25 @@ result = db.query(tree_id, "question", strategy="block", beam_size=3)
 
 Current filesystem benchmark summary lives in [bench/fs_block_beam_vertical.md](bench/fs_block_beam_vertical.md).
 
-Run setup for the snapshot below: `beam_size=3`, `max_turns=10`, `5` filesystem queries on `context7` only.
+Run setup: `fs_query_order=prefix`, `beam_size=3`, `max_turns=10`, `5` filesystem queries on `context7` only.
+
+### Claude Opus 4.6
+
+| Retriever | Avg Time (s) | Avg LLM Calls | Hit@1 | Hit@10 | Total Cost (USD) |
+|---|---:|---:|---:|---:|---:|
+| **Block** | 9.27 | 2.6 | 1.00 | 1.00 | 0.1416 |
+| **Vertical** | 22.85 | 6.8 | 0.40 | 1.00 | 0.1682 |
+| **Beam** | 18.37 | 5.0 | 0.60 | 1.00 | 0.1331 |
 
-### Block vs Beam vs Vertical
+### Claude Sonnet 4.6
 
 | Retriever | Avg Time (s) | Avg LLM Calls | Hit@1 | Hit@10 | Total Cost (USD) |
 |---|---:|---:|---:|---:|---:|
-| **Block** | 5.47 | 1.00 | 1.00 | 1.00 | 0.0762 |
-| **Vertical** | 7.31 | 1.60 | 1.00 | 1.00 | 0.1486 |
-| **Beam** | 20.18 | 4.60 | 0.60 | 0.80 | 0.1328 |
+| **Block** | 7.95 | 2.8 | 1.00 | 1.00 | 0.1670 |
+| **Vertical** | 17.85 | 5.8 | 0.40 | 0.80 | 0.1438 |
+| **Beam** | 17.41 | 4.8 | 0.60 | 1.00 | 0.1338 |
 
-`Block` is the best default on this `context7` snapshot: same retrieval quality as `Vertical`, with lower latency and fewer model calls. `Beam` is still workable, but it trails clearly on retrieval accuracy.
+`Block` is the best default: perfect Hit@1 across both models. `Beam` and `Vertical` are sensitive to model version — `Block` is the most robust choice.
 
 These numbers are benchmark snapshots, not hard guarantees; exact cost and latency will vary with model choice, provider pricing, prompt-cache behavior, and corpus shape.
 

contextdb/config/__init__.py

@@ -24,7 +24,7 @@ def _load_yaml(path: Path) -> dict:
 
 def get_defaults() -> dict:
     """Get default configuration."""
-    return _load_yaml(CONFIG_DIR / "defaults.yaml")
+    return _load_yaml(CONFIG_DIR / "config.yaml")
 
 
 def get_llm_config(provider: str, model: str) -> dict:
@@ -60,12 +60,16 @@ def get_retriever_config(retriever_type: str) -> dict:
 
 
 class Config:
-    """Configuration class for environment variables and LLM client."""
+    """Configuration: .env for keys, config.yaml for settings, env vars override."""
 
+    # Keys — from .env only
     ANTHROPIC_API_KEY = os.getenv("ANTHROPIC_API_KEY")
     OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
-    LLM_PROVIDER = os.getenv("LLM_PROVIDER", "anthropic")
-    LLM_MODEL = os.getenv("LLM_MODEL", "claude-sonnet-4-20250514")
+
+    # Settings — config.yaml, env vars override
+    _cfg = get_defaults().get("llm", {})
+    LLM_PROVIDER = os.getenv("LLM_PROVIDER") or _cfg.get("provider", "anthropic")
+    LLM_MODEL = os.getenv("LLM_MODEL") or _cfg.get("model", "claude-sonnet-4-6")
     DB_PATH = os.getenv("DB_PATH", "context.sqlite")
 
     @classmethod

contextdb/config/defaults.yaml contextdb/config/config.yamlcontextdb/config/defaults.yaml renamed to contextdb/config/config.yaml

@@ -1,12 +1,11 @@
-# Default configuration
+# ConDB configuration
 
 llm:
   provider: anthropic
-  model: claude-sonnet-4-20250514
+  model: claude-sonnet-4-6
   context_limit: 100000
   max_concurrent: 10
 
 retriever:
-  tokens_per_node: 300
   beam_size: 3
   max_turns: 5

contextdb/config/llm/anthropic.yaml

@@ -1,25 +1,13 @@
 # Anthropic models configuration
 
-claude-sonnet-4-20250514:
+claude-opus-4-6:
   context_limit: 200000
   max_concurrent: 50
   rpm_limit: 4000
   tpm_limit: 400000
 
-claude-opus-4-20250514:
+claude-sonnet-4-6:
   context_limit: 200000
   max_concurrent: 50
   rpm_limit: 4000
   tpm_limit: 400000
-
-claude-3-5-sonnet-20241022:
-  context_limit: 200000
-  max_concurrent: 50
-  rpm_limit: 4000
-  tpm_limit: 400000
-
-claude-3-haiku-20240307:
-  context_limit: 200000
-  max_concurrent: 100
-  rpm_limit: 4000
-  tpm_limit: 400000

contextdb/llm.py

@@ -180,7 +180,7 @@ def _build_messages(messages: list[dict], system: str = "") -> list[dict]:
 # ── Default models ────────────────────────────────────────────────
 
 _DEFAULT_MODELS = {
-    "anthropic": "claude-sonnet-4-20250514",
+    "anthropic": "claude-sonnet-4-6",
     "openai": "gpt-4",
 }