How hzi-bifo/ribodetector works
RiboDetector is a command-line rRNA detector for FASTA/FASTQ sequencing reads. Its main workflow loads reads, encodes them, classifies them in GPU or CPU mode, then separates non-rRNA, rRNA, and optionally unclassified reads into output files. Paired-end confidence handling and chunked processing are part of the detector workflow. The excerpts show the GPU model and shared sequence-processing calls; CPU inference uses ONNX Runtime. Some utility and loader abstractions are present, but their runtime connections are not shown.
Main entry points: detect.py GPU detector · detect_cpu.py CPU detector · seq_encoder.py Sequence encoder · parse_config.py Config parser
Components 17
Detection workflow
- GPU detector
ribodetector/detect.py - CPU detector
ribodetector/detect_cpu.py
Read preparation
- FASTX parser
ribodetector/data_loader/fastx_parser.py - Sequence encoder
ribodetector/data_loader/seq_encoder.py - Dataset loader
ribodetector/data_loader/dataset.py - Loader base
ribodetector/base/base_data_loader.py
Classification models
- LSTM model
ribodetector/model/model.py - CPU model
ribodetector/model/model_cpu.py - Model base
ribodetector/base/base_model.py
Configuration and support
- Config parser
ribodetector/parse_config.py - JSON utilities
ribodetector/utils/util.py - Logger helper
ribodetector/logger/logger.py
Other components
- CLI user
- FASTA/FASTQ reads
- Non-rRNA reads
- rRNA reads
- Unclassified reads
Connections 20
- CLI user → GPU detector: runs GPU mode
README.md - CLI user → CPU detector: runs CPU mode
README.md - FASTA/FASTQ reads → GPU detector: provides reads
README.md - FASTA/FASTQ reads → CPU detector: provides reads
README.md - GPU detector → Config parser: uses configuration
ribodetector/detect.py - CPU detector → Config parser: uses configuration
ribodetector/detect_cpu.py - GPU detector → Sequence encoder: encodes reads
ribodetector/detect.py - GPU detector → Dataset loader: loads batches
ribodetector/detect.py - CPU detector → Sequence encoder: encodes reads
ribodetector/detect_cpu.py - Sequence encoder → FASTX parser: parses records
ribodetector/data_loader/seq_encoder.py - GPU detector → LSTM model: classifies reads
ribodetector/detect.py - LSTM model → Model base: inherits from
ribodetector/model/model.py - CPU model → Model base: inherits from
ribodetector/model/model_cpu.py - GPU detector → Non-rRNA reads: writes classified reads
ribodetector/detect.py - GPU detector → rRNA reads: writes classified reads
ribodetector/detect.py - GPU detector → Unclassified reads: writes uncertain reads
ribodetector/detect.py - CPU detector → Non-rRNA reads: writes classified reads
ribodetector/detect_cpu.py - CPU detector → rRNA reads: writes classified reads
ribodetector/detect_cpu.py - CPU detector → Unclassified reads: writes uncertain reads
ribodetector/detect_cpu.py - Config parser → JSON utilities: reads JSON
ribodetector/parse_config.py