Remove LocationParsingConfiguration and LocationPatternMatcher classes, along with related tests and UI components. Update EventOccurrenceParserService to include location validation logic, enhancing warning reporting for long locations and potential date/time patterns. Adjust appsettings and UI to reflect the removal of location parsing settings.

2026-01-09 09:10:22 -05:00
parent 2eae3f205c
commit ea1a4a04ad
9 changed files with 95 additions and 547 deletions
@@ -1,4 +1,5 @@
 using System.Text;
+using System.Text.RegularExpressions;
 using Core.Entities;
 using Core.Models;
 using Core.Parsers;
@@ -93,6 +94,9 @@ public class EventOccurrenceParserService : IEventOccurrenceParserService

                // Copy skipped HS section headers from parser result
                result.SkippedHSSectionHeaders.AddRange(parserResult.SkippedHSSectionHeaders);
+
+                // Validate locations and add warnings for problematic ones
+                ValidateLocations(result);
            }
            finally
            {
@@ -121,5 +125,48 @@ public class EventOccurrenceParserService : IEventOccurrenceParserService

        return result;
    }
+
+    /// <summary>
+    /// Validates locations from parsed occurrences and adds warnings for problematic locations.
+    /// </summary>
+    private static void ValidateLocations(EventOccurrenceParseResult result)
+    {
+        // Collect all unique locations
+        var locations = result.Occurrences.Values
+            .SelectMany(list => list)
+            .Select(eo => eo.Location)
+            .Where(loc => !string.IsNullOrWhiteSpace(loc))
+            .Distinct()
+            .ToList();
+
+        if (!locations.Any())
+            return;
+
+        // Check for long locations (>50 chars)
+        var longLocations = locations.Where(loc => loc != null && loc.Length > 50).ToList();
+        foreach (var loc in longLocations)
+        {
+            if (loc != null)
+            {
+                result.Warnings.Add($"Location '{loc}' is unusually long ({loc.Length} characters) and may contain multiple lines or extra text");
+            }
+        }
+
+        // Check for date/time patterns
+        // Pattern matches: month names with day numbers, time patterns (HH:MM AM/PM), and NOON
+        var dateTimePattern = new Regex(
+            @"\b(January|February|March|April|May|June|July|August|September|October|November|December)\s+\d{1,2}\b|\b\d{1,2}:\d{2}\s*(a|p)\.?m\.?\b|\bNOON\b",
+            RegexOptions.IgnoreCase | RegexOptions.Compiled);
+        
+        var locationsWithDateTime = locations.Where(loc => loc != null && dateTimePattern.IsMatch(loc)).ToList();
+        foreach (var loc in locationsWithDateTime)
+        {
+            if (loc != null)
+            {
+                var match = dateTimePattern.Match(loc);
+                result.Warnings.Add($"Location '{loc}' may contain date/time information: '{match.Value}'");
+            }
+        }
+    }
 }