# fast5

**Repository Path**: mirrors_debian/fast5

## Basic Information

- **Project Name**: fast5
- **Description**: A C++ header-only library for reading Oxford Nanopore Fast5 files
- **Primary Language**: Unknown
- **License**: MIT
- **Default Branch**: master
- **Homepage**: None
- **GVP Project**: No

## Statistics

- **Stars**: 0
- **Forks**: 0
- **Created**: 2020-08-22
- **Last Updated**: 2026-05-17

## Categories & Tags

**Categories**: Uncategorized

**Tags**: None

## README

# -*- mode:org; mode:visual-line; coding:utf-8; -*-

** Fast5 Library

[[http://travis-ci.org/mateidavid/fast5][http://travis-ci.org/mateidavid/fast5.svg?branch=master]] [[https://tldrlegal.com/license/mit-license][http://img.shields.io/:license-mit-blue.svg]]

A lightweight C++11 library to read raw signal data from Oxford Nanopore's Fast5 files.

*** C++

**** Installation

This is a header-only library. You only need to copy [[file:src/fast5.hpp][src/fast5.hpp]] and [[file:src/hdf5_tools.hpp][src/hdf5_tools.hpp]] into your C++ project.

**** Usage

See [[file:src/f5ls.cpp][src/f5ls.cpp]] for an example.

*** Python Wrapper

An optional python wrapper for this library is available through Boost.Python. The wrapper currently implements only read-only access.

**** Installation

#+BEGIN_EXAMPLE
cd python
HDF5_DIR=/usr/local BOOST_DIR=/usr/local make develop-user
#+END_EXAMPLE

Notes:

- HDF5 and Boost.Python must be available, and their locations can be passed on to the Python setup process using the environment variables =HDF5_DIR= and =BOOST_DIR=. Alternatively, the respective include directories, library directories, and library names may be specified explicitly with: =HDF5_INCLUDE_DIR=, =HDF5_LIB_DIR=, =HDF5_LIB=, =BOOST_INCLUDE_DIR=, =BOOST_LIB_DIR=, =BOOST_PYTHON_LIB=. For details, see [[file:python/setup.py][python/setup.py]] and [[file:.travis.yml][.travis.yml]].

- To install =fast5= as a package in a virtualenv, use the target =develop=. To install as a user package, use the target =develop-user=. For details, see [[file:python/Makefile][python/Makefile]].

**** Usage

#+BEGIN_EXAMPLE
import fast5
f = fast5.File("file.000.fast5")
print(f.file_version())
print(f.have_eventdetection_events())
#+END_EXAMPLE

*** f5dump

The program =f5dump= can be used to list and extract some of the contents of =fast5= files, including: raw signals, event-detection events, basecall events, and basecall fastq.

**** Installation

In addition to this =fast5= repository, you will need HDF5 (headers and libraries), as well the the header-only libraries [[https://github.com/mateidavid/tclap.git][TCLAP]] and [[https://github.com/mateidavid/hpptools.git][HPPTOOLS]]. To build =f5dump=, run =make f5dump [VAR1=VALUE1] ...=, where =VAR=-s are used to instruct the [[file:src/Makefile][Makefile]] where to find various dependencies.

**** Usage

In each run, =f5dump= requires exactly one command among: =--ls/--id/--rw/--ed/--ev/--fq=. If no command is given, =--ls= is assumed. It also requires exactly one =fast5= file to inspect.

- In =--ls= mode, =f5dump= lists some of the contents of the file. Sample output:

  #+BEGIN_EXAMPLE
rw      Read_1019
ed      000     Read_1019
ed      001     Read_1019
bc2d    2D_000  2       1       1       1D_000
bc1d    1D_000  0       1       1       001
bc1d    1D_000  1       1       1       001
#+END_EXAMPLE

  Explanations:

  - =rw=: the file contains raw samples from one read, =Read_1019=.

  - =ed=: the file contains 2 event-detection groups, =000= and =001=, both run on raw samples from =Read_1019=.

  - =bc2d=: the file contains 1 basecall group =2D_000= with 2D data (=2=); this group has both fastq data and events (=1 1=); its corresponding 1D basecall group is =1D_000=.

  - =bc1d=: the file contains 1 basecall group =1D_000= with 1D data for each strand (=0= and =1=); each contains fastq data and events (=1 1=); its corresponding event-detection group is =001=.

  Notes:

  - The group names are suffixes understood by the =fast5= library. E.g., the basecall group =RNN_1D_000= would correspond to the HDF5 group =/Analyses/Basecall_RNN_1D_000=.

  - Not all the links between groups are always available. Notably, some =fast5= files are missing the link between a 1D basecall group and its original event-detection group.

- In =--id= mode, =f5dump= dumps =channel_id= and =tracking_id= metadata.

- In =--rw/--ed/--ev/--fq= mode, =f5dump= dumps: raw signal data/event-detection events/basecall events/basecall fastq data.

- Optional selector flags =--gr/--st/--rn= can be used to specify a group name, strand (=0/1/2=), or read name. Not all combinations make sense: e.g, =--st= is ignored for event-detection data.

- Optional output flags =--time-int/--curr-int/--rw-time= can modify the output: convert times into integers, dump raw signal currents in internal integer encoding, and add time stamps to raw signals.

*** License

[[file:LICENSE][MIT License]].