Documentation Index
Fetch the complete documentation index at: https://mintlify.com/kishnahai0806/SteelWorks/llms.txt
Use this file to discover all available pages before exploring further.
SteelWorks exposes two download buttons — one beneath each data table — that let operators and analysts pull the current view into a CSV file without leaving the browser. Both exports are generated on demand from the same service calls that populate the tables, so the CSV always exactly matches what is shown on screen for the active filter selection.
Issue summary CSV
The first download button appears directly below the Issue Summary table:
st.download_button(
label="Download issue summary CSV",
data=service.export_issue_summary_csv(
selection=selection, group_by_line=group_by_line
),
file_name="issue_summary.csv",
mime="text/csv",
)
issue_summary.csv
Columns (ungrouped — group_by_line=False):
| Column | Description |
|---|
week_label | ISO week label (e.g. 2026-W03) |
issue_type_name | Issue category name |
issue_count | Number of issues of that type |
group_by_line=True, the default):
| Column | Description |
|---|
week_label | ISO week label |
line_name | Production line name |
issue_type_name | Issue category name |
issue_count | Number of issues of that type on that line |
Example content (ungrouped)
week_label,issue_type_name,issue_count
2026-W03,material_shortage,1
2026-W03,tool_wear,1
Affected lots CSV
The second download button appears directly below the Affected Lots table:
st.download_button(
label="Download affected lots CSV",
data=service.export_affected_lots_csv(selection=selection),
file_name="affected_lots.csv",
mime="text/csv",
)
affected_lots.csv
Columns:
| Column | Description |
|---|
week_label | ISO week label |
line_name | Production line name |
lot_code | Lot identifier |
issue_count | Number of issues logged against this lot |
issue_types | Comma-separated, sorted list of distinct issue type names |
Example content
week_label,line_name,lot_code,issue_count,issue_types
2026-W03,Line 1,LOT-1001,1,tool_wear
2026-W03,Line 1,LOT-1002,1,material_shortage
Exports reflect the active filter
Both exports are regenerated every time a download button is clicked — they do not cache a previous result. Changing the week selector, deselecting a production line, or toggling the group-by-line checkbox all update what will be downloaded the next time the button is pressed. The data in the file will always match the tables currently visible on screen.
Service methods
Both export methods live on OperationsMetricsService:
service.export_issue_summary_csv(selection=selection, group_by_line=group_by_line)
service.export_affected_lots_csv(selection=selection)
get_* method internally to retrieve the row data, then write it through a csv.DictWriter into an io.StringIO buffer before encoding to bytes:
def _to_csv_bytes(self, rows: list[dict[str, Any]], fieldnames: list[str]) -> bytes:
buffer = io.StringIO()
writer = csv.DictWriter(buffer, fieldnames=fieldnames)
writer.writeheader()
for row in rows:
writer.writerow({field: row.get(field, "") for field in fieldnames})
return buffer.getvalue().encode("utf-8")
bytes encoded as UTF-8, which Streamlit’s st.download_button accepts directly via its data parameter.
The CSV fieldnames are always written in the exact column order shown in the tables above — week_label is always first, issue_count is always last in the summary file. A header row is always present in the output even when the filter scope returns no data rows, so downstream tooling can safely assume a consistent structure regardless of whether the current week has any issues.