seismoai-io-ali 0.1.0

I/O and visualization module for seismic SEG-Y data processing

seismoai — Seismic Data Processing Library

A real Python library for AI-powered seismic data analysis built on the Forge 2D Survey (2017) dataset.

We have Selected these Modules #1 and #2

Package	PyPI name	Description
seismoai_io	seismoai-io	Load SGY files, batch loading, normalization
seismoai_viz	seismoai-viz	Gather images, trace waveforms, frequency spectra

---

---

Main Code that i used Main.py

from seismoai_io import load_sgy, normalize_traces
from seismoai_viz import plot_gather, plot_trace, plot_spectrum

# Load a real SGY file
traces = load_sgy("data/27_1511546140_30100_50100_20171127_150416_752.sgy")
print(traces.shape)   # (167, 4001)

# Normalize
normed = normalize_traces(traces, method="zscore")

# Visualize
fig1 = plot_gather(normed, dt_ms=1.0, title="Record 27 — Forge 2D")
fig1.savefig("gather.png", dpi=150)

fig2 = plot_trace(normed, trace_index=83)
fig2.savefig("trace_83.png", dpi=150)

fig3 = plot_spectrum(traces, trace_index=0, max_freq_hz=200)
fig3.savefig("spectrum.png", dpi=150)

---

seismoai_io API

load_sgy(filepath, *, ignore_geometry=True, dtype=np.float32)

Load a single SEG-Y file → ndarray shape (n_traces, n_samples).

load_sgy_folder(folder, *, extensions=('.sgy','.segy'), verbose=True)

Load all SEG-Y files in a directory → dict[stem → ndarray].

normalize_traces(traces, method='zscore', *, per_trace=True)

Normalize trace amplitudes.

method	Description	Output range
'zscore'	Zero-mean, unit-variance	(−∞, +∞)
'minmax'	Min-Max scaling	[0, 1]
'maxabs'	Maximum absolute value	[−1, 1]

---

seismoai_viz API

plot_gather(traces, *, dt_ms=1.0, title=..., cmap='seismic', clip_percentile=99.0)

2-D colour-fill gather image. Returns Figure.

plot_trace(traces, trace_index=0, *, dt_ms=1.0, title=...)

1-D amplitude-vs-time waveform for one channel. Returns Figure.

plot_spectrum(traces, trace_index=0, *, dt_ms=1.0, max_freq_hz=None, db_scale=True)

FFT amplitude spectrum with Hann window. Returns Figure.

---

Running Tests

pip install pytest
pytest tests/ -v --tb=short

---

Development Setup

git clone https://github.com/Ali1/2826/seismoai.git
cd seismoai
pip install -e ".[dev]"

---

Dataset Used

The SGY files are from the Forge 2D seismic survey (Utah, USA, 2017), recorded with an INOVA Geophysical Hawk system.

• 167 traces per record
• 4001 samples per trace
• 1 ms sample interval (4-second record length)
• IEEE Float32, Little-Endian encoding

---

License

MIT

SeismoAI