Migrating Chinese Stock Data into a Custom Labeling Workflow

Summary

The post documents an attempt to move daily Chinese stock bars into a custom data source and use them as input to an automated labeling module. Its example queries a date-bounded range of stock records, adjusts instrument identifiers, writes the resulting table to a data source, and defines a forward-return label. That label uses a later close relative to the next open, clips extreme values by quantiles, bins the result, and masks cases where the next day’s high and low are equal, described as limit-up sessions.

The evidence is a troubleshooting record rather than a working recipe: the displayed runs include a NameError from referring to an undefined data object and a ValueError caused by ambiguous truth evaluation of a DataFrame. The post does not show a final corrected run or confirm that the labeler accepts the custom data source. Readers should therefore treat the code and error traces as migration clues, check the platform’s expected instrument input type, and verify the behavior of the label expression and data fields before using it for research.

Key ideas

• The example queries daily Chinese stock bars over a specified historical interval and stores them as a custom data source.
• Its label represents a short forward return from the next open to a later close.
• The example clips extreme labels, bins them, and removes observations matching a one-price limit-up condition.
• The recorded attempts fail with an undefined variable and a DataFrame truth-value error.
• The post does not provide a verified fix or a successful end-to-end labeling run.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.