docs(explainer): fix invalid publish_log example in publish-director

Step 6 still showed a legacy publish_log entry with 'metadata' and 'video_path'
fields, which the publish_log schema rejects (entries set additionalProperties:
false). Replace it with the actual schema-valid shape export_bundle returns in
data['publish_log'] (status 'exported', export_path, metadata_used) and note
that it should be persisted directly without extra entry fields.
This commit is contained in:
0xDevNinja
2026-06-30 12:05:23 +05:30
parent 9bcccce2db
commit a41f7fb404

View File

@@ -115,23 +115,23 @@ provider.
### Step 6: Build Publish Log
`export_bundle` already returns a schema-valid `publish_log` in `data["publish_log"]` — persist that directly rather than hand-building one. Do **not** add extra entry fields (the schema sets `additionalProperties: false`; only `platform`, `status`, `url`, `video_id`, `visibility`, `export_path`, `timestamp`, `metadata_used`, `error` are allowed). The shape it returns:
```json
{
"version": "1.0",
"entries": [
{
"platform": "youtube",
"status": "draft",
"timestamp": "2024-01-15T10:30:00Z",
"metadata": {
"status": "exported",
"export_path": "projects/vector-db-explainer/exports",
"timestamp": "2026-01-15T10:30:00+00:00",
"metadata_used": {
"title": "Vector Databases Explained in 60 Seconds",
"description_length": 450,
"tags_count": 8,
"chapters_count": 6,
"thumbnail_ready": false
},
"export_path": "exports/vector-db-explainer/",
"video_path": "renders/output.mp4"
"description": "What vector databases are and when to use them.",
"hashtags": ["#ai", "#vectordb"],
"chapters": [{ "start_seconds": 0, "title": "Introduction" }]
}
}
]
}