Supermind uses an extract-transform-load (ETL) process to ingest data into the Supermind Graph.
The final stage of this process (the "load" stage) involves importing pre-prepared CSV files into the Supermind Graph. These CSV files are referred to as "load jobs".
In production, these "load jobs" will be generated as a result of the previous "extract" and "transform" stages.
This tool allows us to bypass these two stages ("extract" and "transform") by generating mock "load jobs" that can be fed straight into the final "load" stage.
NOTE: a load job actually contains two files... a CSV file for the data, and a JSON file for the metadata.
WARNING #1: Requires 25GB of storage space to run.
WARNING #2: Output folder will contain ~100,000 files. This may crash some IDEs, hence why the output folder isn't a child of this project. We strongly recommend using the terminal to interact with this folder.
To generate the mock "load jobs":
yarn
yarn start
The results will be stored to:
../supermind-load-job-mocks
The process of transforming a document (e.g. webpage, PDF, etc.) into a single standardised format (CSV in our case).
The process of standardising and making sense of data within the extracted table. For example, normalising dates, identifying links between cell values and external tables, etc.
The process of loading a clean, normalised table into Supermind.
A "Supermind Load Job" simply represents data that requires no further processing, but requires inserting into the Supermind Graph.
Superming Load Jobs consist of two files:
-
load-job.json: the metadata file. -
load-job.csv: the data file.
Both files must share the same name (before the extension).