A polished, token-controlled text similarity module based on Transformers
For deploying using docker:
# adapt docker-compose.yaml if needed # inside root directory where docker-compose.yaml is docker-compose -p salted_simcore up -d --build
Visit API doc: http://localhost:9060/docs#/
- open task and obtain token with
/open_taskendpoint - upload your analysis text (e.g. company text regarding sustainability) wit
/upload_text/{token}endpoint - upload your reference text (e.g. sustainability report) wit
/upload_reference/{token}endpoint - choose your analyzer of choice (coarse-performes text based matching, detailed-performs sentence based matching) with
/set_analyzer/{token}endpoint - perform the analysis with
/analyze_project/{token}endpoint - choose your vizualization of choice (heatmap, barplot, scatterplot) with
/set_visualizer/{token}endpoint - perform the vizualization with
/visualize_project/{token}endpoint - list all the task files with
/list_task_files/{token}endpoint - download the results with
/download_file/{token}endpoint using the file name from the list of task files as parameter