Bugfixes and documentation cleanup for v1.0 release.

2026-05-01 11:30:29 -05:00
parent c9e98e14b5
commit f20f06db12
17 changed files with 332 additions and 177 deletions
--- a/schema/full-output.schema.json
+++ b/schema/full-output.schema.json
@@ -0,0 +1,98 @@
+{
+  "$schema": "https://json-schema.org/draft/2020-12/schema",
+  "$id": "https://gitea.maximumdirect.net/eric/seriatim/schema/full-output.schema.json",
+  "title": "seriatim full output transcript",
+  "type": "object",
+  "additionalProperties": false,
+  "required": ["metadata", "segments", "overlap_groups"],
+  "properties": {
+    "metadata": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": [
+        "application",
+        "version",
+        "input_reader",
+        "input_files",
+        "preprocessing_modules",
+        "postprocessing_modules",
+        "output_modules"
+      ],
+      "properties": {
+        "application": { "type": "string" },
+        "version": { "type": "string" },
+        "input_reader": { "type": "string" },
+        "input_files": {
+          "type": "array",
+          "items": { "type": "string" }
+        },
+        "preprocessing_modules": {
+          "type": "array",
+          "items": { "type": "string" }
+        },
+        "postprocessing_modules": {
+          "type": "array",
+          "items": { "type": "string" }
+        },
+        "output_modules": {
+          "type": "array",
+          "items": { "type": "string" }
+        }
+      }
+    },
+    "segments": {
+      "type": "array",
+      "items": { "$ref": "#/$defs/segment" }
+    },
+    "overlap_groups": {
+      "type": "array",
+      "items": { "$ref": "#/$defs/overlap_group" }
+    }
+  },
+  "$defs": {
+    "segment": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": ["id", "source", "speaker", "start", "end", "text"],
+      "properties": {
+        "id": { "type": "integer", "minimum": 1 },
+        "source": { "type": "string" },
+        "source_segment_index": { "type": "integer", "minimum": 0 },
+        "source_ref": { "type": "string" },
+        "derived_from": {
+          "type": "array",
+          "items": { "type": "string" }
+        },
+        "speaker": { "type": "string" },
+        "start": { "type": "number" },
+        "end": { "type": "number" },
+        "text": { "type": "string" },
+        "categories": {
+          "type": "array",
+          "items": { "type": "string" }
+        },
+        "overlap_group_id": { "type": "integer", "minimum": 1 }
+      }
+    },
+    "overlap_group": {
+      "type": "object",
+      "additionalProperties": false,
+      "required": ["id", "start", "end", "segments", "speakers", "class", "resolution"],
+      "properties": {
+        "id": { "type": "integer", "minimum": 1 },
+        "start": { "type": "number" },
+        "end": { "type": "number" },
+        "segments": {
+          "type": "array",
+          "items": { "type": "string" }
+        },
+        "speakers": {
+          "type": "array",
+          "items": { "type": "string" }
+        },
+        "class": { "type": "string" },
+        "resolution": { "type": "string" }
+      }
+    }
+  }
+}