# Segyio

> |

- **Type:** Skill
- **Install:** `agentstack add skill-steadfastasart-geoscience-skills-segyio`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [SteadfastAsArt](https://agentstack.voostack.com/s/steadfastasart)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [SteadfastAsArt](https://github.com/SteadfastAsArt)
- **Source:** https://github.com/SteadfastAsArt/geoscience-skills/tree/main/segyio

## Install

```sh
agentstack add skill-steadfastasart-geoscience-skills-segyio
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# segyio - SEG-Y Seismic Data

## Quick Reference

```python
import segyio

# Read
with segyio.open('seismic.sgy', 'r') as f:
    print(f.tracecount, len(f.samples))
    trace0 = f.trace[0]              # Single trace as numpy array
    data = segyio.tools.collect(f.trace[:])  # All traces

# 3D access (specify inline/xline byte locations)
with segyio.open('seismic.sgy', 'r', iline=189, xline=193) as f:
    inline_100 = f.iline[100]        # 2D array (xlines x samples)
    cube = segyio.tools.cube(f)      # Full 3D cube

# Write
spec = segyio.spec()
spec.samples = samples
spec.tracecount = n_traces
with segyio.create('output.sgy', spec) as f:
    f.trace[0] = data
```

## Key Access Modes

| Mode | Description |
|------|-------------|
| `f.trace[i]` | Sequential trace access by index |
| `f.header[i]` | Trace header access (dict-like) |
| `f.iline[n]` | Inline slice (3D surveys) |
| `f.xline[n]` | Crossline slice (3D surveys) |
| `f.depth_slice[n]` | Horizontal time/depth slice |
| `f.text[0]` | Text header (3200 bytes) |
| `f.bin` | Binary header |

## Essential Operations

### Open and Inspect
```python
with segyio.open('seismic.sgy', 'r') as f:
    print(f"Traces: {f.tracecount}")
    print(f"Samples: {len(f.samples)}")
    print(f"Sample interval: {f.samples[1] - f.samples[0]} ms")
    if f.ilines is not None:
        print(f"Inlines: {f.ilines}")
        print(f"Crosslines: {f.xlines}")
```

### Read Trace Headers
```python
with segyio.open('seismic.sgy', 'r') as f:
    header = f.header[0]
    print(header[segyio.TraceField.INLINE_3D])
    print(header[segyio.TraceField.CDP_X])

    # Get all values for one field
    inlines = f.attributes(segyio.TraceField.INLINE_3D)[:]
```

### Read 3D Data
```python
with segyio.open('seismic.sgy', 'r', iline=189, xline=193) as f:
    inline_data = f.iline[100]       # Shape: (n_xlines, n_samples)
    xline_data = f.xline[200]        # Shape: (n_ilines, n_samples)
    time_slice = f.depth_slice[250]  # Shape: (n_ilines, n_xlines)
    cube = segyio.tools.cube(f)      # Shape: (ilines, xlines, samples)
```

### Read Unstructured Data
```python
# For 2D lines or unsorted data
with segyio.open('seismic.sgy', 'r', ignore_geometry=True) as f:
    data = segyio.tools.collect(f.trace[:])
```

### Create New SEG-Y
```python
import numpy as np

spec = segyio.spec()
spec.samples = np.arange(500) * 4  # 4ms sample rate
spec.tracecount = 100
spec.format = 1  # IBM float

with segyio.create('output.sgy', spec) as f:
    for i in range(100):
        f.trace[i] = data[i]
        f.header[i] = {
            segyio.TraceField.TRACE_SEQUENCE_LINE: i + 1,
        }
```

### Modify Existing File
```python
with segyio.open('seismic.sgy', 'r+') as f:
    f.trace[0] = f.trace[0] * 2.0
    f.header[0][segyio.TraceField.CDP] = 999
```

## Data Formats

| Code | Format |
|------|--------|
| 1 | IBM 4-byte float |
| 2 | 4-byte signed integer |
| 3 | 2-byte signed integer |
| 5 | IEEE 4-byte float |
| 8 | 1-byte signed integer |

## When to Use vs Alternatives

| Tool | Best For |
|------|----------|
| **segyio** | Fast SEG-Y I/O, 3D inline/crossline access, header manipulation |
| **obspy** | Broader seismology: FDSN access, waveform processing, event analysis |
| **segysak** | xarray-based SEG-Y workflows, NetCDF conversion, labeled dimensions |

**Use segyio when** you need fast, low-level access to SEG-Y files -- reading
traces, headers, 3D slices, or creating new SEG-Y files programmatically.

**Use obspy instead** when you need seismological processing beyond file I/O:
instrument response removal, FDSN data fetching, or earthquake analysis.

**Use segysak instead** when you want xarray integration with labeled
dimensions (inline, crossline, time) and easy conversion to NetCDF/Zarr.

## Common Workflows

### Read, inspect, and extract 3D seismic data
```
- [ ] Open file with `segyio.open()`, specify `iline=` and `xline=` byte positions
- [ ] Inspect geometry: trace count, sample count, inline/crossline ranges
- [ ] Read headers to verify coordinate and survey metadata
- [ ] Extract target inline/crossline slices or full cube
- [ ] Apply amplitude scaling or subset extraction as needed
- [ ] Write results to new SEG-Y or export as numpy arrays
```

## Common Issues

| Issue | Solution |
|-------|----------|
| Geometry not detected | Specify `iline=` and `xline=` byte positions |
| Can't read 3D slices | Use `ignore_geometry=True` for unstructured data |
| Wrong inline/xline bytes | Check headers with `f.header[0]` to find correct bytes |
| File not opening | Check file path, permissions, and SEG-Y format validity |

## References

- **[Trace Header Fields](references/header_fields.md)** - Complete header field reference (bytes, types)
- **[Troubleshooting](references/troubleshooting.md)** - Common problems and solutions

## Scripts

- **[scripts/inspect_segy.py](scripts/inspect_segy.py)** - Inspect SEG-Y file structure and geometry
- **[scripts/extract_subset.py](scripts/extract_subset.py)** - Extract subset of traces or inlines

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [SteadfastAsArt](https://github.com/SteadfastAsArt)
- **Source:** [SteadfastAsArt/geoscience-skills](https://github.com/SteadfastAsArt/geoscience-skills)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-steadfastasart-geoscience-skills-segyio
- Seller: https://agentstack.voostack.com/s/steadfastasart
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.