podfish 0.1.0

Ultralight local and remote POD5 connector

podfish

Fish reads directly from POD5 files located in public cloud bucket without downloading the whole object.

Quick start

pip install podfish

Fish a single read:

import numpy as np
import matplotlib.pyplot as plt
import podfish as pf

example_pod5 = "https://s3.amazonaws.com/foo/bar.pod5"
selected_read_id = "12345-67890"

with pf.Reader(example_pod5) as reader:
    read = next(reader.reads(selection=[selected_read_id]))
    sample_rate = read.run_info.sample_rate
    signal = read.signal
    time = np.arange(len(signal)) / sample_rate
    plt.plot(time, signal)

Fish the first ten reads:

with pf.Reader(example_pod5) as reader:
    for count, read in enumerate(reader.reads(limit=10), start=1):
        signal = read.signal
        time = np.arange(len(signal)) / read.run_info.sample_rate

        print(count, read.read_id)
        plt.figure(figsize=(4, 2))
        plt.plot(time, signal)
        plt.show()
        plt.close()

Acknowledgement

• Brandon Saint-John's pod5-rs