com.blockether.vis.internal.image-convert — com.blockether/vis 0.1.13

Container conversion for attachment payloads -- NOT optimization.

vis stores and replays attachment bytes VERBATIM: whatever the user dropped is what is stored and what the provider sees. The one thing that cannot be passed through is a container no vision wire accepts (attachments/ provider-image-media-types): a BMP is a perfectly good image and an SVG is a perfectly good figure, and both are a hard 400 on every provider. Since an attachment REPLAYS on every later turn, one such row kills the whole session.

So this namespace does exactly one thing: turn bytes a provider REFUSES into the same picture in a container it ACCEPTS.

• raster (BMP, or anything else ImageIO can decode) -> PNG, 1:1, lossless
• vector (.svg / gzipped .svgz) -> rendered PNG

No downscaling, no quality knob, no size heuristics -- image optimization was removed deliberately and is not hiding here. Every failure mode (undecodable payload, missing AWT/ImageIO stack, malformed markup) returns nil, never throws; the caller then SKIPS the attachment rather than shipping bytes the provider refuses.

*enabled?* clj

Kill switch -- bind false to make conversion unavailable (what a build with no AWT/ImageIO stack looks like).

rasterize-svg clj

(rasterize-svg data)

(rasterize-svg data {:keys [max-dimension]})

Render an SVG (.svg or gzipped .svgz) into a PNG every vision wire accepts. Returns {:bytes :media-type :size :original-size :width :height :original-width :original-height}, or nil when the document does not parse, AWT/ImageIO is missing (GraalVM native-image on macOS) or conversion is disabled. Never throws.

The ASPECT RATIO is the whole game: a squashed chart is a wrong chart, and the model cannot tell the difference. The raster size is the document's own resolved size -- percentages resolved against the viewBox, not jsvg's raw .size, which reports width="100%" as 100px -- and the only scaling that ever happens is the svg-max-raster-dimension ceiling, one factor applied to both edges.

The canvas is filled WHITE first: SVG figures routinely leave the background transparent, and a transparent PNG of black axes is what a viewer -- or a model -- sees as an empty image.

svg-max-raster-dimension clj

Hard ceiling on the rendered long edge. A vector document can declare any size at all, and a 40000px canvas is an OOM, not a picture.

svg-media-type? clj

(svg-media-type? media-type)

svg-media-types clj

Vector containers vis RASTERIZES on the way in instead of refusing. No provider reads SVG -- it is markup, not pixels -- but unlike a HEIC the pixels are recoverable: the document is rendered here, so a matplotlib savefig(format="svg") or a dropped icon still reaches the model.

to-provider-safe clj

(to-provider-safe data media-type)

Re-container data into something every vision wire accepts (PNG). Payloads whose container is already wire-safe come back UNTOUCHED, byte for byte -- this is a validity fix, never a size one. Returns the same map as rasterize-svg, or nil when the payload can be turned into pixels by neither ImageIO nor the SVG renderer (HEIC, AVIF) or conversion is disabled -- the caller must then skip it.

wire-safe-media-types clj

Containers that already need NO conversion (mirrors attachments/provider-image-media-types). Kept here so to-provider-safe can hand these back byte-identical instead of decoding and re-encoding a payload that was already fine.