viewinline 0.3.1__tar.gz → 0.4.0__tar.gz

viewinline-0.4.0/CITATION.cff

@@ -0,0 +1,44 @@
+cff-version: 1.2.0
+message: "If you use this software, please cite it using the metadata from this file."
+title: "viewinline"
+abstract: >-
+  A tiny, non-interactive command-line viewer that displays rasters/photos,
+  vectors, and tabular data (CSV/Parquet) inline in the terminal — no GUI,
+  no X11, and no file downloads. Designed for quick visual inspection at the
+  command line, and especially useful on HPC systems and remote servers over
+  SSH, where images render on the local terminal.
+type: software
+authors:
+  - given-names: Keiko
+    family-names: Nomura
+    # orcid: "https://orcid.org/0000-0000-0000-0000"   # optional: add if you have one
+  - given-names: Filip
+    family-names: Kral
+    # orcid: "https://orcid.org/0000-0000-0000-0000"   # optional: add if you have one
+repository-code: "https://github.com/nkeikon/inlineviewer"
+url: "https://github.com/nkeikon/inlineviewer"
+license: Apache-2.0
+version: 0.3.2
+date-released: 2026-06-06
+# ---------------------------------------------------------------------------
+# The DOI below is the VERSION DOI for v0.3.2 (a specific archived snapshot).
+# To make citations always resolve to your LATEST release, replace it with the
+# Zenodo CONCEPT / "all versions" DOI, shown on the record under "Cite all
+# versions". The version DOI stays useful in `identifiers` for reproducibility.
+# ---------------------------------------------------------------------------
+doi: "10.5281/zenodo.20564266"
+identifiers:
+  - type: doi
+    value: "10.5281/zenodo.20564266"
+    description: "Zenodo archive of version v0.3.2"
+keywords:
+  - geospatial
+  - raster
+  - vector
+  - terminal
+  - command-line
+  - visualization
+  - GeoTIFF
+  - NetCDF
+  - GeoParquet
+  - remote-sensing

{viewinline-0.3.1 → viewinline-0.4.0}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.4
+Metadata-Version: 2.5
 Name: viewinline
-Version: 0.3.1
+Version: 0.4.0
 Summary: Quick look geospatial viewer for the terminal, with inline image previews
 Project-URL: Homepage, https://github.com/nkeikon/viewinline
 Project-URL: Repository, https://github.com/nkeikon/viewinline
@@ -81,6 +81,11 @@ viewinline hyperspectral.nc --subset 22 --bands 10-54 --gallery 5x11
 viewinline path/to/vector.geojson
 viewinline boundaries.geoparquet --color-by population --colormap viridis
 
+# Save to file
+viewinline path/to/file.tif --export out.png   # save the rendered image as PNG (or .jpg)
+viewinline scene.tif --rgb 4 3 2 --export rgb.png
+viewinline dem.tif --colormap terrain --display 1 --export dem.png   # full-res with colormap
+
 # CSV and Parquet
 viewinline data.csv                          # preview rows and columns
 viewinline data.parquet --describe           # summary statistics
@@ -134,6 +139,7 @@ You can also force the chafa path on any terminal by setting `INLINE_VIEWER_ENGI
 - **NetCDF/HDF Support:** Display variables from NetCDF (.nc) and HDF5 (.h5, .hdf5) files with automatic nodata detection and multi-slice navigation
 - **Parquet/GeoParquet:** Render GeoParquet as vector maps or view as tabular data
 - **Tabular View for Vectors:** Use `--table` to access CSV-style operations (filter, sort, describe, hist) on any vector file  
+- **AI-agent inspection:** `--info` returns file metadata and statistics as JSON; `--export` saves a quick-look image for an agent to inspect. Designed for AI coding-agent workflows.
 
 ## Supported formats  
 **Rasters**  
@@ -180,6 +186,7 @@ You can also force the chafa path on any terminal by setting `INLINE_VIEWER_ENGI
 - Variables with 4+ dimensions are not supported
 - For a complete variable list, use `ncdump -h file.nc` or `viewtif`
 
+
 ## Dependencies
 
 **Core dependencies** (installed automatically):
@@ -213,16 +220,18 @@ You can also force the chafa path on any terminal by setting `INLINE_VIEWER_ENGI
 ```
 General:
   --display DISPLAY     Resize only the displayed image (0.5=smaller, 2=bigger). Default: auto-fit to terminal.
+  --info                Print file metadata and statistics as JSON, then exit (agent-facing). For NetCDF/HDF, lists variables; combine with --subset N to inspect one. Always returns JSON, including on error.
+  --export PATH         Save the rendered image to PATH (.png/.jpg) and open it. Works with any display flag. Prints {"path": "..."} as the final line.
 
 Raster:
   --band BAND           Band number to display (single raster), or slice number for NetCDF. (default: 1)
   --bands BANDS         Display multiple bands as a grid. Accepts ranges (30-40), lists (3,4,5), or mixed (1,5,10-15).
+  --rgb R G B           Three band numbers for RGB display (e.g., --rgb 4 3 2). Overrides default 1 2 3.
+  --rgbfiles R G B      Three single-band rasters for RGB composite. Can also provide as positional arguments.
   --timestep INTEGER    Alias for --band when working with NetCDF files.
   --subset INTEGER      Variable index for NetCDF/HDF files (e.g., --subset 1).
   --reduce DIM_NAME     For 3D NetCDF variables, specify which dimension to use as the band/slider axis.  Auto-detected if omitted.
   --colormap            Apply colormap to single-band rasters. Flag without the color scheme → 'terrain'.
-  --rgb R G B           Three band numbers for RGB display (e.g., --rgb 4 3 2). Overrides default 1 2 3.
-  --rgbfiles R G B      Three single-band rasters for RGB composite. Can also provide as positional arguments.
   --vmin VMIN           Minimum pixel value for raster display scaling.
   --vmax VMAX           Maximum pixel value for raster display scaling.
   --nodata NODATA       Override nodata value for rasters if dataset metadata is missing or incorrect.
@@ -249,6 +258,95 @@ CSV and Parquet:
   --select COLUMNS      Select specific columns (space separated). Example: --select Country City
   --sql QUERY           Execute full DuckDB SQL query. Use 'data' as table name. Example: --sql "SELECT * FROM data WHERE Poverty > 40"
 ```
+## AI-agent inspection (`--info`, `--export`)
+
+An optional command aimed at AI coding-agent workflows (Claude Code, Codex, Cursor, and similar). An agent can confirm its code *ran*, but not whether the geospatial file it produced is *sensible* — wrong CRS, unexpected dimensions, all-NoData, NaN/Inf values, or a flipped output. `--info` answers **"what did I create?"** as machine-readable JSON.
+
+```bash
+viewinline result.tif --info                 # metadata + statistics as JSON, then exit
+```
+
+It reports facts, not judgments: format, dimensions, bands, dtype, CRS, resolution, bounds, nodata, and per-band statistics. For NetCDF/HDF it lists variables; add `--subset N` to inspect one:
+
+```bash
+viewinline data.nc --info                    # list variables
+viewinline data.nc --subset 7 --info         # inspect variable 7
+viewinline scene.hdf --subset 1 --info       # inspect HDF subdataset 1
+```
+
+`--info` always returns JSON, including a `{"readable": false, "error": ...}` object on unreadable input, so an agent can always parse the result. Pairs well with `--export` (see Usage) when the agent wants to *see* the output too, not just read its metadata.
+
+### `--info` — structured inspection
+
+Prints file metadata as JSON and exits (no image is drawn). Reports facts, not judgments — the agent interprets them in context.
+
+```bash
+viewinline result.tif --info
+```
+```json
+{
+  "format": "GTiff",
+  "filename": "result.tif",
+  "dimensions": [1001, 1001],
+  "bands": 3,
+  "dtype": "uint16",
+  "crs": "EPSG:32631",
+  "resolution": [10.0, 10.0],
+  "bounds": [590520.0, 5780620.0, 600530.0, 5790630.0],
+  "nodata": null,
+  "statistics": {
+    "method": "full",
+    "bands_total": 3,
+    "bands_reported": 3,
+    "per_band": [
+      {"band": 1, "min": 0.0, "max": 10964.0, "mean": 1009.32, "valid_fraction": 1.0, "naninf_fraction": 0.0}
+    ]
+  }
+}
+```
+
+For **NetCDF and HDF**, `--info` lists the file's variables/subdatasets; add `--subset N` to inspect one:
+
+```bash
+viewinline data.nc --info                    # list variables
+viewinline data.nc --subset 7 --info         # inspect variable 7 (dims, dtype, units, stats)
+viewinline scene.hdf --subset 1 --info       # inspect HDF subdataset 1
+```
+
+`--info` **always returns JSON**, including on failure. Unsupported or unreadable inputs return a structured error rather than crashing, so an agent can always parse the result:
+
+```json
+{"error": "no directly-readable bands (file has subdatasets)", "readable": false, "subdataset_count": 22}
+```
+
+Notes on the output:
+- `crs` is an `EPSG:code` when one can be resolved, the full WKT string when the CRS has no EPSG code (e.g. MODIS Sinusoidal), or `null` when the file has no CRS.
+- Statistics exclude NoData and non-finite pixels; `valid_fraction` and `naninf_fraction` report how much was excluded.
+- For files with many bands, per-band stats are capped (a `bands_reported` < `bands_total` and a `note` indicate truncation — absence of a band's stats does **not** imply a problem).
+- Large rasters are sampled for statistics (`"method": "sampled"`); small ones use every pixel (`"method": "full"`).
+
+### `--export` — visual inspection
+
+Saves the rendered image to a file (PNG or JPEG, chosen by extension) so an agent can also inspect it with its vision capabilities, then prints the path as JSON:
+
+```bash
+viewinline result.tif --export out.png
+```
+```json
+{"path": "/abs/path/out.png"}
+```
+
+`--export` works with **any display option** — the saved image is exactly what viewinline would render, after all flags are applied:
+
+```bash
+viewinline scene.tif --rgb 4 3 2 --export rgb.png
+viewinline dem.tif --colormap terrain --export dem.png
+viewinline result.tif --display 1 --export full_res.png   # full resolution instead of terminal-fit
+```
+
+The exported image is a quick-look representation for catching problems metadata can't reveal — blank output, stripes, artifacts, holes, wrong orientation, unexpected extent, or bad color scaling — **not** a publication-quality rendering.
+
+> **Scope:** `--info` and `--export` tell you what a file *is* and what it *looks like*. Neither claims the scientific result is *correct* — that judgment stays with the agent.
 
 ## Need help?
 NASA staff can ask questions about usage via the documentation-based assistant 'viewtif + viewgeom + viewinline Helper' via the ChatGSFC Agent Marketplace.

{viewinline-0.3.1 → viewinline-0.4.0}/README.md

@@ -48,6 +48,11 @@ viewinline hyperspectral.nc --subset 22 --bands 10-54 --gallery 5x11
 viewinline path/to/vector.geojson
 viewinline boundaries.geoparquet --color-by population --colormap viridis
 
+# Save to file
+viewinline path/to/file.tif --export out.png   # save the rendered image as PNG (or .jpg)
+viewinline scene.tif --rgb 4 3 2 --export rgb.png
+viewinline dem.tif --colormap terrain --display 1 --export dem.png   # full-res with colormap
+
 # CSV and Parquet
 viewinline data.csv                          # preview rows and columns
 viewinline data.parquet --describe           # summary statistics
@@ -101,6 +106,7 @@ You can also force the chafa path on any terminal by setting `INLINE_VIEWER_ENGI
 - **NetCDF/HDF Support:** Display variables from NetCDF (.nc) and HDF5 (.h5, .hdf5) files with automatic nodata detection and multi-slice navigation
 - **Parquet/GeoParquet:** Render GeoParquet as vector maps or view as tabular data
 - **Tabular View for Vectors:** Use `--table` to access CSV-style operations (filter, sort, describe, hist) on any vector file  
+- **AI-agent inspection:** `--info` returns file metadata and statistics as JSON; `--export` saves a quick-look image for an agent to inspect. Designed for AI coding-agent workflows.
 
 ## Supported formats  
 **Rasters**  
@@ -147,6 +153,7 @@ You can also force the chafa path on any terminal by setting `INLINE_VIEWER_ENGI
 - Variables with 4+ dimensions are not supported
 - For a complete variable list, use `ncdump -h file.nc` or `viewtif`
 
+
 ## Dependencies
 
 **Core dependencies** (installed automatically):
@@ -180,16 +187,18 @@ You can also force the chafa path on any terminal by setting `INLINE_VIEWER_ENGI
 ```
 General:
   --display DISPLAY     Resize only the displayed image (0.5=smaller, 2=bigger). Default: auto-fit to terminal.
+  --info                Print file metadata and statistics as JSON, then exit (agent-facing). For NetCDF/HDF, lists variables; combine with --subset N to inspect one. Always returns JSON, including on error.
+  --export PATH         Save the rendered image to PATH (.png/.jpg) and open it. Works with any display flag. Prints {"path": "..."} as the final line.
 
 Raster:
   --band BAND           Band number to display (single raster), or slice number for NetCDF. (default: 1)
   --bands BANDS         Display multiple bands as a grid. Accepts ranges (30-40), lists (3,4,5), or mixed (1,5,10-15).
+  --rgb R G B           Three band numbers for RGB display (e.g., --rgb 4 3 2). Overrides default 1 2 3.
+  --rgbfiles R G B      Three single-band rasters for RGB composite. Can also provide as positional arguments.
   --timestep INTEGER    Alias for --band when working with NetCDF files.
   --subset INTEGER      Variable index for NetCDF/HDF files (e.g., --subset 1).
   --reduce DIM_NAME     For 3D NetCDF variables, specify which dimension to use as the band/slider axis.  Auto-detected if omitted.
   --colormap            Apply colormap to single-band rasters. Flag without the color scheme → 'terrain'.
-  --rgb R G B           Three band numbers for RGB display (e.g., --rgb 4 3 2). Overrides default 1 2 3.
-  --rgbfiles R G B      Three single-band rasters for RGB composite. Can also provide as positional arguments.
   --vmin VMIN           Minimum pixel value for raster display scaling.
   --vmax VMAX           Maximum pixel value for raster display scaling.
   --nodata NODATA       Override nodata value for rasters if dataset metadata is missing or incorrect.
@@ -216,6 +225,95 @@ CSV and Parquet:
   --select COLUMNS      Select specific columns (space separated). Example: --select Country City
   --sql QUERY           Execute full DuckDB SQL query. Use 'data' as table name. Example: --sql "SELECT * FROM data WHERE Poverty > 40"
 ```
+## AI-agent inspection (`--info`, `--export`)
+
+An optional command aimed at AI coding-agent workflows (Claude Code, Codex, Cursor, and similar). An agent can confirm its code *ran*, but not whether the geospatial file it produced is *sensible* — wrong CRS, unexpected dimensions, all-NoData, NaN/Inf values, or a flipped output. `--info` answers **"what did I create?"** as machine-readable JSON.
+
+```bash
+viewinline result.tif --info                 # metadata + statistics as JSON, then exit
+```
+
+It reports facts, not judgments: format, dimensions, bands, dtype, CRS, resolution, bounds, nodata, and per-band statistics. For NetCDF/HDF it lists variables; add `--subset N` to inspect one:
+
+```bash
+viewinline data.nc --info                    # list variables
+viewinline data.nc --subset 7 --info         # inspect variable 7
+viewinline scene.hdf --subset 1 --info       # inspect HDF subdataset 1
+```
+
+`--info` always returns JSON, including a `{"readable": false, "error": ...}` object on unreadable input, so an agent can always parse the result. Pairs well with `--export` (see Usage) when the agent wants to *see* the output too, not just read its metadata.
+
+### `--info` — structured inspection
+
+Prints file metadata as JSON and exits (no image is drawn). Reports facts, not judgments — the agent interprets them in context.
+
+```bash
+viewinline result.tif --info
+```
+```json
+{
+  "format": "GTiff",
+  "filename": "result.tif",
+  "dimensions": [1001, 1001],
+  "bands": 3,
+  "dtype": "uint16",
+  "crs": "EPSG:32631",
+  "resolution": [10.0, 10.0],
+  "bounds": [590520.0, 5780620.0, 600530.0, 5790630.0],
+  "nodata": null,
+  "statistics": {
+    "method": "full",
+    "bands_total": 3,
+    "bands_reported": 3,
+    "per_band": [
+      {"band": 1, "min": 0.0, "max": 10964.0, "mean": 1009.32, "valid_fraction": 1.0, "naninf_fraction": 0.0}
+    ]
+  }
+}
+```
+
+For **NetCDF and HDF**, `--info` lists the file's variables/subdatasets; add `--subset N` to inspect one:
+
+```bash
+viewinline data.nc --info                    # list variables
+viewinline data.nc --subset 7 --info         # inspect variable 7 (dims, dtype, units, stats)
+viewinline scene.hdf --subset 1 --info       # inspect HDF subdataset 1
+```
+
+`--info` **always returns JSON**, including on failure. Unsupported or unreadable inputs return a structured error rather than crashing, so an agent can always parse the result:
+
+```json
+{"error": "no directly-readable bands (file has subdatasets)", "readable": false, "subdataset_count": 22}
+```
+
+Notes on the output:
+- `crs` is an `EPSG:code` when one can be resolved, the full WKT string when the CRS has no EPSG code (e.g. MODIS Sinusoidal), or `null` when the file has no CRS.
+- Statistics exclude NoData and non-finite pixels; `valid_fraction` and `naninf_fraction` report how much was excluded.
+- For files with many bands, per-band stats are capped (a `bands_reported` < `bands_total` and a `note` indicate truncation — absence of a band's stats does **not** imply a problem).
+- Large rasters are sampled for statistics (`"method": "sampled"`); small ones use every pixel (`"method": "full"`).
+
+### `--export` — visual inspection
+
+Saves the rendered image to a file (PNG or JPEG, chosen by extension) so an agent can also inspect it with its vision capabilities, then prints the path as JSON:
+
+```bash
+viewinline result.tif --export out.png
+```
+```json
+{"path": "/abs/path/out.png"}
+```
+
+`--export` works with **any display option** — the saved image is exactly what viewinline would render, after all flags are applied:
+
+```bash
+viewinline scene.tif --rgb 4 3 2 --export rgb.png
+viewinline dem.tif --colormap terrain --export dem.png
+viewinline result.tif --display 1 --export full_res.png   # full resolution instead of terminal-fit
+```
+
+The exported image is a quick-look representation for catching problems metadata can't reveal — blank output, stripes, artifacts, holes, wrong orientation, unexpected extent, or bad color scaling — **not** a publication-quality rendering.
+
+> **Scope:** `--info` and `--export` tell you what a file *is* and what it *looks like*. Neither claims the scientific result is *correct* — that judgment stays with the agent.
 
 ## Need help?
 NASA staff can ask questions about usage via the documentation-based assistant 'viewtif + viewgeom + viewinline Helper' via the ChatGSFC Agent Marketplace.

viewinline-0.4.0/SKILL.md

@@ -0,0 +1,199 @@
+---
+name: viewinline
+description: Terminal viewer for rasters, vectors, and tabular data. Use for quick visual inspection without leaving the shell—preview images in a folder gallery, inspect geospatial files after GDAL workflows, explore CSV data with histograms/scatter plots, or verify files before commit. Provides machine-readable `--info` (metadata + statistics as JSON) and `--export` (save the rendered image) for programmatically checking generated geospatial files. Non-interactive; everything controlled via command-line flags.
+tags: [visualization, terminal, raster, vector, csv, geospatial, gdal, ml, inspection, workflow]
+---
+
+# Viewinline Skill
+
+Quick-look terminal viewer for geospatial and tabular data. Think of it as `ls` for visual files—designed for command-line workflows where you want to inspect data without leaving the terminal or opening a separate application.
+
+Renders natively in iTerm2, WezTerm, Konsole, Rio, and Contour. Falls back to `chafa` (ASCII art with 24-bit color) in other terminals like kitty, Ghostty, Terminal.app, VS Code, and Linux terminals. Works over SSH without X11 forwarding or VNC.
+
+## When to Use
+
+- **File browsing**: `find . -name "*.jpg" | xargs -n1 viewinline` to visually scan results
+- **Gallery view**: `viewinline path/to/folder --gallery 4x3` to preview all images in a directory
+- **Band inspection**: Display specific bands of a raster as a grid: `--bands 10-50` or `--bands 11,15,30,45`
+- **Geospatial workflows**: Inspect raster/vector outputs after GDAL transformations
+- **ML training data**: Review training image galleries without switching windows
+- **Data exploration**: Quick CSV visualizations (histograms, scatter plots, summary stats)
+- **Pre-commit checks**: Verify images, maps, and data files before pushing
+- **Remote servers**: Works over SSH from your local terminal
+- **Verify generated files (agent workflows)**: After writing a GeoTIFF/NetCDF/vector with GDAL, rasterio, or xarray, run `viewinline FILE --info` to check the result is sensible (correct CRS, dimensions, value range, not all-NoData/NaN) as JSON, and `viewinline FILE --export out.png` to visually inspect it
+
+## Supported Formats
+
+**Rasters:** GeoTIFF, PNG, JPEG, NetCDF, HDF5, HDF4  
+**Vectors:** GeoJSON, Shapefile, GeoPackage, GeoParquet  
+**Tabular:** CSV, Parquet (with `pyarrow`)
+
+## Inspecting Files (JSON output)
+
+For programmatic checks, two flags return machine-readable output instead of drawing an image. Useful after generating a file to confirm it's sensible before continuing.
+
+- `--info` — print file metadata and statistics as JSON, then exit. Reports facts, not judgments (the caller interprets them).
+  - Rasters (GeoTIFF, NetCDF, HDF): format, dimensions, bands, dtype, CRS, resolution, bounds, nodata, per-band stats (min/max/mean, valid_fraction, naninf_fraction).
+  - Vectors (GeoJSON, Shapefile, GeoPackage, GeoParquet): feature count, geometry type, CRS, bounds, columns.
+  - For NetCDF/HDF: `--info` lists variables/subdatasets; add `--subset N` to inspect one.
+  - Always returns JSON, including `{"readable": false, "error": ...}` on unreadable input — safe to parse in all cases.
+- `--export PATH` — save the rendered image to PATH (`.png`/`.jpg`) and print `{"path": "..."}`. Works with any display flag (`--rgb`, `--colormap`, `--band`, `--display`), saving exactly what would be drawn.
+
+```bash
+# check a raster you just wrote
+viewinline result.tif --info
+
+# check a vector
+viewinline boundaries.geojson --info
+
+# NetCDF: list variables, then inspect one
+viewinline data.nc --info
+viewinline data.nc --subset 7 --info
+
+# save a quick-look image to inspect visually
+viewinline result.tif --export out.png
+viewinline scene.tif --rgb 4 3 2 --export rgb.png
+```
+
+Note: `--info` and `--export` report what a file *is* and what it *looks like* — not whether the scientific result is correct.
+
+## Core Flags
+
+**Inspection (JSON output, no image):**
+- `--info` — print metadata + statistics as JSON, then exit (rasters and vectors; lists variables for NetCDF/HDF with `--subset N`)
+- `--export PATH` — save the rendered image to PATH (`.png`/`.jpg`); prints `{"path": "..."}`
+
+**Raster Display:**
+- `--rgb R G B` — Specify band order for RGB (e.g., `--rgb 4 3 2`)
+- `--rgbfiles R.tif G.tif B.tif` — Create RGB composite from separate files
+- `--band N` — Display specific band or NetCDF timestep
+- `--bands RANGE` — Display multiple bands as a grid (e.g., `10-50` or `11,15,30,45`)
+- `--gallery NxM` — Show all images in folder as thumbnails (e.g., `4x3` grid)
+- `--colormap NAME` — Apply colormap (e.g., `plasma`, `viridis`, `terrain`)
+- `--vmin VAL --vmax VAL` — Set min/max for scaling
+- `--display SCALE` — Resize displayed output (0.5=smaller, 2=bigger)
+
+**Vector Display:**
+- `--color-by COLUMN` — Color features by attribute value
+- `--colormap NAME` — Apply colormap to coloring
+- `--width W` — Line width for boundaries
+- `--edgecolor COLOR` — Edge color (hex or named)
+- `--table` — Display as tabular data instead of rendering geometry
+
+**Tabular Data (CSV/Parquet):**
+- `--describe [COL]` — Summary statistics for all numeric columns or one
+- `--hist [COL]` — Histograms for numeric columns
+- `--scatter X Y` — Scatter plot of two columns
+- `--where EXPR` — Filter rows (DuckDB required): `--where "year > 2010"`
+- `--sort COL` — Sort by column (ascending); use `--desc` for descending
+- `--limit N` — Limit output rows
+- `--select COL1 COL2` — Choose specific columns to display
+- `--sql QUERY` — Full DuckDB SQL (use `data` as table name)
+
+**NetCDF/HDF:**
+- `--subset N` — Select variable by index
+- `--band N` or `--timestep N` — Select slice along time/band axis
+- `--reduce DIM_NAME` — Override auto-detected band axis for non-standard dimensions
+
+## Examples
+
+### Image browsing
+```bash
+# Browse all JPGs in directory
+find . -name "*.jpg" | xargs viewinline
+
+# Gallery of all images in a folder (4 columns × 3 rows)
+viewinline path/to/images --gallery 4x3
+```
+
+### Raster workflows
+```bash
+# Display single band
+viewinline data.tif --band 2
+
+# RGB composite from three bands
+viewinline multiband.tif --rgb 4 3 2
+
+# Create RGB from separate files
+viewinline R.tif G.tif B.tif
+
+# Display band range as gallery
+viewinline hyperspectral.tif --bands 10-50 --gallery 5x5
+
+# Apply colormap with min/max scaling
+viewinline temp.nc --subset 1 --colormap plasma --vmin 273 --vmax 310
+```
+
+### GDAL verification
+```bash
+# Reproject and check result
+gdalwarp -t_srs EPSG:3857 input.tif temp.tif
+viewinline temp.tif --colormap terrain
+
+# Resample and verify
+gdal_translate -outsize 50% 50% temp.tif output.tif
+viewinline output.tif
+```
+
+### Vector inspection
+```bash
+# View vector file
+viewinline boundaries.geojson
+
+# Color features by attribute, with colormap
+viewinline boundaries.geoparquet --color-by population --colormap viridis
+
+# View vector as tabular data
+viewinline counties.shp --table
+viewinline data.geoparquet --table --where "POP > 100000" --sort POP --desc
+```
+
+### Data exploration
+```bash
+# Preview CSV
+viewinline data.csv
+
+# Summary statistics
+viewinline data.parquet --describe
+
+# Histograms for all numeric columns
+viewinline data.csv --hist
+
+# Scatter plot
+viewinline data.csv --scatter area_km2 year
+
+# Filter and sort
+viewinline data.csv --where "year > 2010" --sort population --desc
+
+# SQL query
+viewinline data.csv --sql "SELECT * FROM data WHERE area > 100 ORDER BY year"
+
+# Batch CSV inspection with xargs (use -I {} to control argument order)
+find . -name "*.csv" | xargs -n1 -I {} viewinline {} --describe
+```
+
+### Verify generated files (agent workflows)
+```bash
+# Reproject, then check the result is sensible as JSON
+gdalwarp -t_srs EPSG:3857 input.tif output.tif
+viewinline output.tif --info
+
+# Confirm a written raster isn't all-NoData / has expected CRS and range
+viewinline result.tif --info
+
+# List a NetCDF's variables, then inspect the data variable
+viewinline output.nc --info
+viewinline output.nc --subset 7 --info
+
+# Save a quick-look PNG to inspect visually
+viewinline result.tif --export check.png
+```
+
+## Tips
+
+- Use `--gallery` to quickly preview batches of training images or search results
+- Pipe with `xargs`: `find . -newer file.txt | xargs viewinline` to inspect recent changes
+- Add aliases: `alias check='viewinline --describe --hist'`
+- Works in tmux if outer terminal is iTerm2; falls back to ASCII art in other terminals
+- Install `chafa` for better coverage: `brew install chafa` (macOS), `apt install chafa` (Linux)
+- SSH-friendly: images render on your local terminal, not the remote server

viewinline-0.4.0/context7.json

@@ -0,0 +1,20 @@
+```json
+{
+  "$schema": "https://context7.com/schema/context7.json",
+  "projectTitle": "viewinline",
+  "description": "Terminal inline geospatial preview tool for rasters and vectors over SSH and HPC workflows, with machine-readable --info and --export for AI-agent file inspection.",
+  "excludeFolders": [
+    "tests",
+    "examples/large-data",
+    "build",
+    "dist"
+  ],
+  "rules": [
+    "Prefer SSH and terminal-based workflows in examples.",
+    "Use inline terminal image rendering instead of GUI workflows.",
+    "Highlight support for GeoTIFF, NetCDF, GeoJSON, and GeoParquet.",
+    "Emphasize compatibility limitations for terminal graphics protocols.",
+    "For programmatic or AI-agent use, note --info (JSON metadata) and --export (save rendered image)."
+  ]
+}
+```