Index elastiflix-reloaded
Load data
Add movies to the catalog one at a time, or import a file to add them (append) or to replace the whole catalog (bulk). The three plot vectors (plot_256, plot_512, plot_1024) are generated automatically from plot with jina-embeddings-v5-omni-small.
Load mode
With bulk, the current catalog (6,899 movies) is deleted at publication and only the movies in this file remain.
JSON file
Upload process
- Browser check · pendingExtension, type and size (25 MB max) before anything is uploaded.
- Upload to the server · pendingThe application server receives the file and repeats the type and size checks.
- Validation against the mapping · pendingThe type of every field in every record; if anything doesn't match, everything stops.
- Write to the catalog · pendingRecords are added to the current index through the alias, in batches of 100; if an id already exists, the movie is replaced.
- Publication · pendingThe elastiflix-reloaded alias switches to the new index and the previous index is deleted, in a single operation.
This file can't be loaded
The file was rejected in the browser before upload, so it never left your computer and nothing was imported.
Before upload, the browser checks the extension and type (.json or .csv, no double extensions), the size (25 MB max) and that the content is UTF-8 text. The server repeats these checks when it receives the file.
Column mapping
| CSV column | Index field | Cell content | Status |
|---|
Validation against the index mapping
Records read
0
Valid
0
Will replace existing
0
Errors
0
The upload stopped: the data doesn't match the index mapping
The whole file was checked and 0 errors were found. No records were imported. Fix the file and upload it again; the upload only continues when the whole file matches the mapping.
| Record | id | Field | Received | The mapping expects |
|---|
Showing 0 of 0 errors.
Download full error report (.csv)Checks passed
Warnings · they don't stop the upload
Rules: every field must exist in the index mapping and have its type (release_date as yyyy-MM-dd, adult as a boolean, keywords and belongs_to_collection as nested objects); id, title and plot are required, and id can't repeat. The whole file is validated before anything is written: if a single record doesn't match, the upload stops and no records are imported. In append mode, if an id already exists in the index, the record replaces the existing movie; in bulk mode, the catalog keeps only the movies in the file.