Refine manual deployment guide

MANUAL_TESTING.md

@@ -11,7 +11,7 @@ cd /home/sourav/Documents/blahblah/cicd-debugger-env
 source .venv/bin/activate
 ```
 
-Always run tests with the project interpreter below, not system `python3`:
+Always run commands with the project interpreter below, not system `python3`:
 
 ```bash
 .venv/bin/python
@@ -20,8 +20,14 @@ Always run tests with the project interpreter below, not system `python3`:
 If dependencies are missing:
 
 ```bash
-pip install -r requirements.txt
-pip install openenv-core
+.venv/bin/python -m pip install -r requirements.txt
+```
+
+Optional sanity checks:
+
+```bash
+which python
+.venv/bin/python -c "import fastapi,openai; print('venv deps ok')"
 ```
 
 ## 1) Run All Unit Tests
@@ -64,7 +70,7 @@ Expected output should include lines starting with:
 Start server (Terminal 1):
 
 ```bash
-uvicorn server.app:app --host 0.0.0.0 --port 7860
+.venv/bin/python -m uvicorn server.app:app --host 0.0.0.0 --port 7860
 ```
 
 Test endpoints (Terminal 2):
@@ -76,6 +82,11 @@ curl -sS http://127.0.0.1:7860/state
 curl -sS -X POST http://127.0.0.1:7860/step -H 'Content-Type: application/json' -d '{"action":"read_logs: inspect failing stage logs"}'
 ```
 
+Expected `POST /reset` response includes typed reward fields:
+- `reward`
+- `reward_model.value`
+- `reward_model.components`
+
 ## 5) Docker Manual Test
 
 Build image:
@@ -124,10 +135,16 @@ bash validate-submission.sh https://your-space-name.hf.space .
 
 ## 8) Baseline Across Easy/Medium/Hard
 
-Run the baseline runner:
+OpenAI client mode:
+
+```bash
+OPENAI_API_KEY="<your_openai_compatible_api_key>" .venv/bin/python baseline_inference.py --max-steps 5 --policy-mode imp --trajectories 3 --force-local-env
+```
+
+Offline reproducible mode:
 
 ```bash
-.venv/bin/python baseline_inference.py --max-steps 5 --policy-mode imp --trajectories 3 --force-local-env
+.venv/bin/python baseline_inference.py --max-steps 5 --policy-mode imp --trajectories 3 --offline --force-local-env
 ```
 
 This writes:
@@ -136,6 +153,17 @@ This writes:
 artifacts/baseline_scores.json
 ```
 
+## 9) Quick Full Verification
+
+Run this sequence for a full local confidence check:
+
+```bash
+.venv/bin/python -m unittest discover -s tests -v
+.venv/bin/python -m openenv.cli.__main__ validate
+.venv/bin/python baseline_inference.py --max-steps 5 --policy-mode imp --trajectories 3 --offline --force-local-env
+docker build -t cicd-debugger-env .
+```
+
 ## Common Issues
 
 - `pytest: command not found`:
@@ -143,6 +171,10 @@ artifacts/baseline_scores.json
 - `ModuleNotFoundError: No module named 'fastapi'` or `No module named 'openai'`:
   - You are likely using system `python3` (`/usr/bin/python3`).
   - Re-run with `.venv/bin/python -m unittest discover -s tests -v`.
+- `pip install -r requirements.txt` exits with code 1:
+  - Use `.venv/bin/python -m pip install -r requirements.txt`.
+  - Confirm venv is active: `source .venv/bin/activate`.
+  - If network is restricted, retry later or use a network that can reach PyPI.
 - `Bind for 0.0.0.0:7860 failed: port is already allocated`:
   - Use host port 7861 (or another free port).
 - OpenAI/HF credential issues during online inference: