Files
chapter-organizer/tools/GoogleSheetsScheduleImport/ScheduleGridExtractor.cs
poprhythm f400813667 Update EventOccurrence parsing to use EventOccurrenceParseGroup for improved data structure
This commit refactors the EventOccurrence parsing logic to utilize the EventOccurrenceParseGroup class, enhancing the organization of parsed occurrences by grouping them based on event definitions and optional section levels. The changes include updates to the EventOccurrenceParseResult, EventOccurrenceParser, and EventOccurrenceParserService to accommodate the new grouping structure. Additionally, tests are modified to reflect these changes, ensuring that the parsing functionality remains intact and accurate. This refactor improves data handling and aligns with the overall architecture of the application.
2026-04-04 21:55:44 -04:00

150 lines
5.0 KiB
C#

namespace GoogleSheetsScheduleImport;
/// <summary>
/// Interprets location headers + time rows + colored/value blocks as occurrence lines.
/// </summary>
public static class ScheduleGridExtractor
{
public static List<ParsedOccurrenceLine> Extract(
GridSheetModel grid,
string month,
int day,
List<string> warnings)
{
var result = new List<ParsedOccurrenceLine>();
if (grid.RowCount < 2 || grid.ColCount < 2)
{
warnings.Add($"Sheet '{grid.SheetTitle}': grid too small; need at least a header row and time column.");
return result;
}
var locations = new string[grid.ColCount];
for (var c = 1; c < grid.ColCount; c++)
{
var header = grid.Values[0][c];
locations[c] = string.IsNullOrWhiteSpace(header) ? $"Column {c + 1}" : header.Trim();
}
var rowTimes = new TimeOnly?[grid.RowCount];
rowTimes[0] = null;
for (var r = 1; r < grid.RowCount; r++)
{
var cell = grid.Values[r][0];
if (!TimeCellParser.TryParse(cell, out var t))
{
if (!string.IsNullOrWhiteSpace(cell))
warnings.Add($"Sheet '{grid.SheetTitle}' row {r + 1}: could not parse time label '{cell}'.");
rowTimes[r] = null;
}
else
rowTimes[r] = t;
}
var slot = InferSlotDuration(rowTimes, warnings, grid.SheetTitle);
for (var c = 1; c < grid.ColCount; c++)
{
var location = locations[c];
var r = 1;
while (r < grid.RowCount)
{
if (!IsOccupied(grid, r, c))
{
r++;
continue;
}
var startRow = r;
var name = grid.Values[r][c] ?? "";
var key = BlockKey(grid, r, c);
var endRow = r;
while (endRow + 1 < grid.RowCount &&
IsOccupied(grid, endRow + 1, c) &&
BlockKey(grid, endRow + 1, c) == key)
{
endRow++;
}
var startTime = rowTimes[startRow];
if (startTime == null)
{
warnings.Add(
$"Sheet '{grid.SheetTitle}' ({location}): block at row {startRow + 1} has no parseable start time in column A.");
r = endRow + 1;
continue;
}
var endInstant = ComputeEndTime(rowTimes, endRow, grid.RowCount, slot);
var timeRange = TimeFormatter.ToParserTimeRange(startTime.Value, endInstant);
var title = name.Trim();
if (string.IsNullOrEmpty(title))
title = "(no title)";
result.Add(new ParsedOccurrenceLine(
Name: title,
Month: month,
Day: day,
TimeRange: timeRange,
Location: location,
SourceRowStart: startRow,
SourceRowEnd: endRow,
SourceCol: c));
r = endRow + 1;
}
}
return result;
}
private static bool IsOccupied(GridSheetModel grid, int r, int c)
{
var v = grid.Values[r][c];
var bg = grid.BackgroundKeys?[r][c];
if (!string.IsNullOrWhiteSpace(v))
return true;
return bg != null; // colored empty cell
}
private static (string NameKey, string? Bg) BlockKey(GridSheetModel grid, int r, int c)
{
var raw = grid.Values[r][c] ?? "";
var v = raw.Trim();
var bg = grid.BackgroundKeys?[r][c];
return (v, bg);
}
private static TimeOnly ComputeEndTime(TimeOnly?[] rowTimes, int endRow, int rowCount, TimeSpan slot)
{
if (endRow + 1 < rowCount && rowTimes[endRow + 1] != null)
return rowTimes[endRow + 1]!.Value;
var lastStart = rowTimes[endRow] ?? throw new InvalidOperationException();
return lastStart.Add(slot);
}
private static TimeSpan InferSlotDuration(TimeOnly?[] rowTimes, List<string> warnings, string sheetTitle)
{
var deltas = new List<int>();
for (var i = 1; i < rowTimes.Length - 1; i++)
{
if (rowTimes[i] == null || rowTimes[i + 1] == null)
continue;
var minutes = (int)(rowTimes[i + 1]!.Value - rowTimes[i]!.Value).TotalMinutes;
if (minutes > 0 && minutes <= 24 * 60)
deltas.Add(minutes);
}
if (deltas.Count == 0)
{
warnings.Add($"Sheet '{sheetTitle}': could not infer time-slot length from column A; defaulting to 30 minutes.");
return TimeSpan.FromMinutes(30);
}
var g = deltas.GroupBy(d => d).OrderByDescending(g => g.Count()).First();
return TimeSpan.FromMinutes(g.Key);
}
}