IO Module (`stx.io`)

Universal file I/O through stx.io.save() and stx.io.load(). Format is determined by file extension.

Quick Reference

import scitex as stx

# Save
stx.io.save(data, "output.csv")       # DataFrame → CSV
stx.io.save(arr, "output.npy")        # ndarray → NumPy
stx.io.save(fig, "output.png")        # Figure → PNG + CSV + YAML recipe
stx.io.save(obj, "output.pkl")        # any object → pickle
stx.io.save(d, "output.yaml")         # dict → YAML
stx.io.save(d, "output.json")         # dict → JSON

# Load
data = stx.io.load("input.csv")       # → DataFrame
arr = stx.io.load("input.npy")        # → ndarray
obj = stx.io.load("input.pkl")        # → original object
d = stx.io.load("input.yaml")         # → dict

Supported Formats

Extension	Save Type	Load Return
`.csv`	DataFrame, dict, ndarray	DataFrame
`.npy`	ndarray	ndarray
`.npz`	dict of ndarrays	NpzFile
`.pkl` / `.pickle`	any Python object	original object
`.yaml` / `.yml`	dict	dict
`.json`	dict, list	dict or list
`.png` / `.jpg` / `.svg` / `.pdf` / `.tiff`	matplotlib Figure	ndarray (image) or Figure
`.hdf5` / `.h5`	dict, ndarray	dict or ndarray
`.mat`	dict	dict
`.pth` / `.pt`	PyTorch state_dict	dict
`.parquet`	DataFrame	DataFrame
`.txt`	str	str

Figure Saving

When saving a matplotlib Figure, stx.io.save automatically:

Saves the image file (.png, .svg, etc.)
Exports a .csv with the plotted data extracted from axes
Exports a .yaml recipe for reproducing the figure

fig, ax = stx.plt.subplots()
ax.plot_line(x, y, label="signal")
stx.io.save(fig, "results/plot.png")
# Creates:
#   results/plot.png       (the figure)
#   results/plot.csv       (x, y data)
#   results/plot.yaml      (recipe for stx.plt.reproduce)

Provenance Tracking

Inside @stx.session, every save and load records file hashes to a local SQLite database for reproducibility verification.

@stx.session
def main(logger=stx.INJECTED, **kw):
    data = stx.io.load("input.csv")    # hash recorded as input
    stx.io.save(result, "output.csv")   # hash recorded as output

Other Functions

stx.io.glob(pattern) – Find files matching a glob pattern
stx.io.load_configs(config_dir) – Load and merge YAML config files

API Reference

See scitex.io API Reference for the auto-generated Python API.