Skip to content

gap_plan.md

Latest commit

 

History

History
119 lines (91 loc) · 4.89 KB

gap_plan.md

File metadata and controls

119 lines (91 loc) · 4.89 KB

Gap Analysis: wmt_requirements.txt vs post-init_requirements.txt

Overview

Direct comparison of desired packages in wmt_requirements.txt with actual runtime environment in post-init_requirements.txt (pip freeze output after extension initialization).

Last Updated: 2025-07-29

Executive Summary

  • All required packages are present - No missing packages
  • 96.7% package coverage (58 of 60 active packages, 2 intentionally commented out)
  • Major version gaps exist for several critical packages due to Python 3.8 constraints
  • 4 additional packages in post-init that are not in wmt requirements (likely dependencies)

Package Comparison Analysis

Missing Packages

None - All packages from wmt_requirements.txt are present in post-init_requirements.txt

Additional Packages (in post-init but not in wmt)

  • importlib-resources==6.4.5 - Backport for resource access
  • nose==1.3.7 - Testing framework (likely a dependency)
  • PyWavelets==1.4.1 - Wavelet transforms (scikit-image dependency)
  • zipp==3.20.2 - ZIP archive utilities (backport dependency)

Commented Out Packages (intentionally excluded)

  • onnxruntime==1.20.0 - ONNX model runtime
  • opencv-python==4.11.0.86 - Using native python3-opencv from SDK instead
  • ultralytics==8.3.4 - YOLO framework
  • ultralytics-thop==2.0.14 - Model complexity analysis

Version Comparison Table

Critical Version Differences (2+ major versions behind)

Package Desired (wmt) Actual (post-init) Gap Impact
pandas 2.3.0 1.0.5 1.2.5 major versions Missing modern data analysis features
numpy 2.3.0 1.24.4 ~1 major version Missing latest performance optimizations
protobuf 6.31.1 3.20.3 3 major versions Compatibility issues with modern protocols
pytz 2025.2 2021.3 4 years Outdated timezone database
tifffile 2025.6.11 2023.7.10 2 years Missing recent TIFF format support
tzlocal 5.3.1 2.1 3 major versions Limited timezone detection

Significant Version Differences (1 major version behind)

Package Desired (wmt) Actual (post-init) Notes
matplotlib 3.10.3 3.7.5 Missing latest plotting features
networkx 3.5 3.1 Graph algorithm improvements
pillow 11.2.1 10.4.0 Image processing enhancements
seaborn 0.13.2 0.11.2 Statistical visualization updates
watchdog 6.0.0 4.0.2 File system monitoring improvements

Minor Version Differences (<1 major version)

Package Desired (wmt) Actual (post-init)
APScheduler 3.11.0 3.10.1
contourpy 1.3.2 1.1.1
filelock 3.18.0 3.16.1
fonttools 4.58.2 4.57.0
fsspec 2025.5.1 2025.3.0
imageio 2.37.0 2.35.1
kiwisolver 1.4.8 1.4.7
MarkupSafe 3.0.2 2.1.5
pyparsing 3.2.3 3.1.4
python-dateutil 2.9.0.post0 2.8.2
ruamel.yaml 0.18.14 0.17.21
ruamel.yaml.clib 0.2.12 0.2.7
scikit-image 0.24.0 0.21.0
scipy 1.15.3 1.10.1
termcolor 2.5.0 2.4.0
torch 2.5.1 2.4.1
torchvision 0.20.1 0.19.1
typing_extensions 4.14.0 4.13.2
urllib3 2.4.0 2.2.3

Special Cases

  • rknn-toolkit-lite2:
    • wmt: Specified as local wheel file path
    • post-init: Version 2.3.2 (successfully installed)

Python 3.8 Compatibility Impact

Many desired package versions require Python 3.9+ or even 3.10+:

  • numpy 2.3.0: Requires Python ≥3.10
  • pandas 2.3.0: Requires Python ≥3.9
  • matplotlib 3.10.3: Requires Python ≥3.9
  • pillow 11.2.1: Requires Python ≥3.9
  • scikit-image 0.24.0: Requires Python ≥3.10
  • tifffile 2025.6.11: Requires Python ≥3.10

Recommendations

High Priority Updates

  1. pandas: Critical for data analysis - explore backporting features or upgrading Python
  2. numpy: Foundation package - performance and compatibility concerns
  3. protobuf: Severe version gap - may cause protocol compatibility issues

Medium Priority Updates

  1. matplotlib: Visualization capabilities limited by older version
  2. scikit-image: Advanced image processing features missing
  3. pillow: Image format support may be limited

Low Priority Updates

  1. Minor version differences generally don't impact functionality
  2. Additional packages in post-init are beneficial dependencies

Conclusion

While all required packages are installed, significant version gaps exist primarily due to Python 3.8 constraints. The environment is functional but lacks modern features and optimizations available in newer package versions. Consider:

  1. Upgrading to Python 3.9+ to access modern package versions
  2. Accepting current limitations with documented workarounds
  3. Selectively updating critical packages through custom builds