Types¶
geoembed.core.types
¶
Core type definitions for the geoembed pipeline.
ChipMetadata(chip_id, parent_path, col_off, row_off, width, height)
dataclass
¶
Metadata for a single image chip (virtual — no pixel data).
This is the standard input format for the embedding pipeline. It contains enough information to locate and read the chip from its parent COG.
Attributes:
| Name | Type | Description |
|---|---|---|
chip_id |
str
|
Unique identifier for this chip. |
parent_path |
str
|
Path to the parent COG image. |
col_off |
int
|
Column offset (pixels) within the parent image. |
row_off |
int
|
Row offset (pixels) within the parent image. |
width |
int
|
Chip width in pixels. |
height |
int
|
Chip height in pixels. |
EmbeddingResult(chip_id, embedding)
dataclass
¶
Output of the embedding pipeline for a single chip.
Attributes:
| Name | Type | Description |
|---|---|---|
chip_id |
str
|
The chip this embedding was generated from. |
embedding |
list[float]
|
The embedding vector (typically 768-dim for ViT-Base models). |