fix: enable unbuffered Python output in Docker

[mkhnsn]

Summary

• Adds ENV PYTHONUNBUFFERED=1 to the Dockerfile so container logs appear in real-time
• Without this, Python fully buffers stdout when not connected to a TTY, causing docker logs -f to show nothing until the container exits

Depends on

• Active thermal monitoring, boundary checks, and merged community fixes #9 (this branch is rebased on top of PR Active thermal monitoring, boundary checks, and merged community fixes #9)

Test plan

• Build the Docker image and run the container
• Verify docker logs -f <container> shows output in real-time during benchmarking

🤖 Generated with Claude Code

Summary by CodeRabbit

• New Features

  • Added "--set-values" apply-only mode and richer CLI controls (IP, max-temp, output-dir).
  • Expanded output metrics: hashrate standard deviation, average power, and average fan speed.
  • Results files now include timestamped filenames and selectable output directory.

• Bug Fixes / Reliability

  • Improved runtime stability and safer benchmarking with restart/retry handling and better stabilization checks.

• Documentation

  • Major README overhaul with detailed CLI, Docker, configuration, and usage examples.

• Chores

  • Docker: streaming logs enabled and default /results directory exposed as a volume.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds an unbuffered Python setting and a default /results volume in the Dockerfile, expands README with CLI/Docker usage and details, and significantly enhances bitaxe_hashrate_benchmark.py with a new --set-values mode, richer metrics (hashrate stddev, power, fan speed), dynamic timestamped result filenames, and stabilization/retry logic.

Changes

Cohort / File(s)	Summary
Docker Configuration Dockerfile	Adds PYTHONUNBUFFERED=1, creates /results directory, marks it as a volume, and updates the entrypoint to pass --output-dir /results.
Documentation README.md	Comprehensive rewrite: expanded CLI/docs, Docker usage examples, new --set-values documentation, revised defaults (timings, steps, limits), and extended output/configuration/safety/benchmarking guidance.
Benchmark Application bitaxe_hashrate_benchmark.py	Large functional changes: new --set-values mode, custom help formatter, dynamic result_filename() with timestamp suffix, running_stddev() streaming metric, extended benchmark_iteration() returns (adds hashrateStdDev, averagePower, averageFanSpeed), set_system_settings()/restart_system() now return status and include stabilization monitoring and retry-on-overheat logic, enhanced result payloads and save behavior, and richer runtime guards and status output.

Sequence Diagrams

sequenceDiagram
    actor User
    participant CLI as CLI Parser
    participant Device as Bitaxe Device
    participant FS as Result File

    User->>CLI: Run with --set-values + voltage/frequency
    CLI->>CLI: Validate inputs
    CLI->>Device: set_system_settings(voltage, frequency)
    Device->>Device: Apply settings & restart
    Device-->>CLI: Return status (success/fail)
    CLI->>FS: Optionally log/apply result_filename()
    CLI->>User: Exit with status

Loading

sequenceDiagram
    actor User
    participant Bench as Benchmark Loop
    participant Device as Bitaxe Device
    participant Monitor as Stabilization Monitor
    participant Results as Result Storage

    User->>Bench: Start benchmark run
    Bench->>Device: set_system_settings(voltage, frequency)
    Bench->>Device: restart_system()
    Device-->>Monitor: Emit streaming metrics (hashrate, temp, power, fan)
    Monitor->>Monitor: Evaluate stabilization (variance, temp, retries)
    alt Stabilized
        Bench->>Bench: Compute running_stddev, averages
        Bench->>Results: Save JSON via result_filename()
    else Overheat / Unstable
        Bench->>Bench: Adjust voltage/frequency and retry
        Bench->>Device: Re-apply adjusted settings
    end
    Bench-->>User: Return enriched result payload

Loading

Estimated code review effort

🎯 4 (Complex) | ⏱️ ~60 minutes

Poem

🐰 In my burrow I tweak volts and tune,
I watch stddev shimmer beneath the moon,
Docker unbuffered, logs streaming bright,
Results stamped with timestamps through the night,
I hop with joy — power, fan, and hash take flight!

🚥 Pre-merge checks | ✅ 2 | ❌ 1

❌ Failed checks (1 warning)

Check name	Status	Explanation	Resolution
Docstring Coverage	⚠️ Warning	Docstring coverage is 0.00% which is insufficient. The required threshold is 80.00%.	Write docstrings for the functions missing them to satisfy the coverage threshold.

✅ Passed checks (2 passed)

Check name	Status	Explanation
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title 'fix: enable unbuffered Python output in Docker' is specific and clearly related to the main Dockerfile change, but the PR also includes significant changes to README.md and bitaxe_hashrate_benchmark.py introducing new features (--output-dir flag, standardization metrics, enhanced CLI controls).

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

Actionable comments posted: 3

🤖 Fix all issues with AI agents

In `@bitaxe_hashrate_benchmark.py`:
- Around line 46-58: The help/epilog text in the ArgumentParser (where parser =
argparse.ArgumentParser(...)) contains a typo in the script name
("bitaxe_hasrate_benchmark.py" missing the "h"); update all example lines in the
epilog that reference that filename to the correct
"bitaxe_hashrate_benchmark.py" so the usage examples match the actual script
name and avoid user confusion.
- Around line 339-363: The stabilization loop that calls get_system_info()
currently swallows all exceptions and can treat repeated telemetry failures as a
successful stabilization; update the loop in bitaxe_hashrate_benchmark.py to
track consecutive read errors (e.g., add a failure_counter initialized before
the while, increment on exception or when info is falsy, and reset on a
successful read), log the failure reason via print or a logger when an error
occurs, and if the failure_counter exceeds a small threshold (e.g., 3
consecutive failures) exit the function with False to fail stabilization; keep
the existing temperature/power/vr checks (temp, vr_temp, power) and only reset
the counter when a valid info dict is returned.

In `@README.md`:
- Around line 132-133: Update the README description of the benchmark output
filename to match the actual behavior: change the static name
"bitaxe_benchmark_results_<ip_address>.json" to the runtime filename pattern
that includes the timestamp/suffix (e.g.,
"bitaxe_benchmark_results_<ip_address>_<timestamp>.json" or similar), and ensure
any surrounding text or examples reference that updated filename format so docs
match the code that writes the timestamped files.

[coderabbitai]

Actionable comments posted: 0

Caution

Some comments are outside the diff and can’t be posted inline due to platform limitations.

⚠️ Outside diff range comments (1)

bitaxe_hashrate_benchmark.py (1)

508-513: ⚠️ Potential issue | 🟠 Major

Guard VR-temp trimming to avoid ZeroDivisionError on short series.

trimmed_vr_temps can be empty when fewer than 7 VR temp samples are available, causing division by zero. Consider trimming only when enough samples exist, otherwise average the available values or skip VR temp.

🛠️ Suggested fix

-        if vr_temps:
-            sorted_vr_temps = sorted(vr_temps)
-            trimmed_vr_temps = sorted_vr_temps[6:]  # Remove first 6 elements only
-            average_vr_temp = sum(trimmed_vr_temps) / len(trimmed_vr_temps)
+        if vr_temps:
+            sorted_vr_temps = sorted(vr_temps)
+            if len(sorted_vr_temps) > 6:
+                trimmed_vr_temps = sorted_vr_temps[6:]  # Remove first 6 elements only
+            else:
+                trimmed_vr_temps = sorted_vr_temps
+            average_vr_temp = sum(trimmed_vr_temps) / len(trimmed_vr_temps)