peter/RealCV
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

feat: Reduce false positives in company verification

Browse Source 
Major improvements to company name matching accuracy:

- Add well-known brands dictionary with correct Companies House numbers
  for fast-track verification (Boots, Legal & General, EY, etc.)
- Add safe expansion words (UK, LIMITED, GROUP, PLC) that don't change
  company identity
- Fix core word validation to require original company's core words
- Remove overly aggressive skip words that removed meaningful identifiers
  (industries, technology, consulting, services, etc.)
- Add industry context hints for AI matching
- Fix CVBatchTester JSON deserialization for test files

Before: 98% verified but with false positives like:
- Boots → BOOTS AND BEARDS (wrong)
- Legal & General → LEGAL LIMITED (wrong)

After: 97% verified with correct matches:
- Boots → BOOTS UK LIMITED (correct)
- Legal & General → fast-tracked to correct company

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
Peter Foster
2026-01-22 19:01:11 +00:00
parent 94ca6e1b9a
commit 3d666d5f9c
7 changed files with 814 additions and 114 deletions
Download Patch File Download Diff File Expand all files Collapse all files

249
tools/CVBatchTester/Program.cs
View File

@@ -1,4 +1,5 @@
using System.Text.Json;
using System.Text.Json.Serialization;
using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.Configuration;
using Microsoft.Extensions.DependencyInjection;
@@ -12,23 +13,86 @@ using RealCV.Infrastructure.Services;
namespace CVBatchTester;
// DTOs for test JSON format (snake_case with nested personal object)
record TestCVData
{
public string? CvId { get; init; }
public string? Category { get; init; }
public List<string>? ExpectedFlags { get; init; }
public TestPersonalData? Personal { get; init; }
public string? Profile { get; init; }
public List<TestEmploymentEntry>? Employment { get; init; }
public List<TestEducationEntry>? Education { get; init; }
public List<string>? Skills { get; init; }
}
record TestPersonalData
{
public string? Name { get; init; }
public string? Email { get; init; }
public string? Phone { get; init; }
public string? Address { get; init; }
public string? LinkedIn { get; init; }
}
record TestEmploymentEntry
{
public string? Company { get; init; }
public string? JobTitle { get; init; }
public string? StartDate { get; init; }
public string? EndDate { get; init; }
public string? Location { get; init; }
public string? Description { get; init; }
public List<string>? Achievements { get; init; }
}
record TestEducationEntry
{
public string? Institution { get; init; }
public string? Qualification { get; init; }
public string? Subject { get; init; }
public string? Classification { get; init; }
public string? StartDate { get; init; }
public string? EndDate { get; init; }
}
class Program
{
private static StreamWriter? _logWriter;
private static readonly JsonSerializerOptions JsonOptions = new()
{
PropertyNameCaseInsensitive = true,
PropertyNamingPolicy = JsonNamingPolicy.SnakeCaseLower,
Converters = { new JsonStringEnumConverter() }
};
static async Task<int> Main(string[] args)
{
var folderPath = args.FirstOrDefault() ?? AskForFolder();
if (string.IsNullOrEmpty(folderPath) || !Directory.Exists(folderPath))
{
Console.WriteLine($"Error: Folder not found: {folderPath}");
Console.WriteLine("Usage: CVBatchTester <folder-path>");
Console.WriteLine(" e.g. CVBatchTester /home/user/cvs");
Log($"Error: Folder not found: {folderPath}");
Log("Usage: CVBatchTester <folder-path> [--output <file>]");
Log(" e.g. CVBatchTester /home/user/cvs");
Log(" e.g. CVBatchTester /home/user/cvs --output /tmp/results.log");
return 1;
}
Console.WriteLine($"CV Batch Verification Tester");
Console.WriteLine($"Processing CVs from: {folderPath}");
Console.WriteLine(new string('=', 80));
// Check for --output flag
var outputIndex = Array.IndexOf(args, "--output");
var logPath = outputIndex >= 0 && outputIndex < args.Length - 1
? args[outputIndex + 1]
: Path.Combine(folderPath, $"batch-results-{DateTime.Now:yyyyMMdd-HHmmss}.log");
_logWriter = new StreamWriter(logPath, false) { AutoFlush = true };
Log($"CV Batch Verification Tester");
Log($"Processing CVs from: {folderPath}");
Log($"Output log: {logPath}");
Log($"Started: {DateTime.Now:yyyy-MM-dd HH:mm:ss}");
Log(new string('=', 80));
// Setup DI
var services = new ServiceCollection();
@@ -39,15 +103,16 @@ class Program
var cvFiles = Directory.GetFiles(folderPath, "*.*", SearchOption.TopDirectoryOnly)
.Where(f => f.EndsWith(".pdf", StringComparison.OrdinalIgnoreCase) ||
f.EndsWith(".docx", StringComparison.OrdinalIgnoreCase) ||
f.EndsWith(".doc", StringComparison.OrdinalIgnoreCase))
f.EndsWith(".doc", StringComparison.OrdinalIgnoreCase) ||
f.EndsWith(".json", StringComparison.OrdinalIgnoreCase))
.OrderBy(f => f)
.ToList();
Console.WriteLine($"Found {cvFiles.Count} CV files\n");
Log($"Found {cvFiles.Count} CV files\n");
if (cvFiles.Count == 0)
{
Console.WriteLine("No CV files found (.pdf, .docx, .doc)");
Log("No CV files found (.pdf, .docx, .doc, .json)");
return 1;
}
@@ -63,9 +128,9 @@ class Program
foreach (var cvFile in cvFiles)
{
Console.WriteLine($"\n{new string('=', 80)}");
Console.WriteLine($"[{++processedCount}/{cvFiles.Count}] {Path.GetFileName(cvFile)}");
Console.WriteLine(new string('=', 80));
Log($"\n{new string('=', 80)}");
Log($"[{++processedCount}/{cvFiles.Count}] {Path.GetFileName(cvFile)}");
Log(new string('=', 80));
try
{
@@ -74,17 +139,30 @@ class Program
var companyVerifier = scope.ServiceProvider.GetRequiredService<ICompanyVerifierService>();
var eduVerifier = scope.ServiceProvider.GetRequiredService<IEducationVerifierService>();
// Parse CV
await using var stream = File.OpenRead(cvFile);
var cv = await parser.ParseAsync(stream, Path.GetFileName(cvFile));
// Parse CV - handle JSON files differently
CVData cv;
if (cvFile.EndsWith(".json", StringComparison.OrdinalIgnoreCase))
{
var jsonContent = await File.ReadAllTextAsync(cvFile);
var testCv = JsonSerializer.Deserialize<TestCVData>(jsonContent, JsonOptions)
?? throw new InvalidOperationException("Failed to deserialize JSON CV");
Console.WriteLine($"Candidate: {cv.FullName}");
// Convert TestCVData to CVData
cv = ConvertTestCVData(testCv);
Log($"Loaded JSON CV: {cv.FullName}");
}
else
{
await using var stream = File.OpenRead(cvFile);
cv = await parser.ParseAsync(stream, Path.GetFileName(cvFile));
Log($"Parsed CV: {cv.FullName}");
}
// Verify Employers
if (cv.Employment?.Count > 0)
{
Console.WriteLine($"\nEMPLOYERS ({cv.Employment.Count}):");
Console.WriteLine(new string('-', 60));
Log($"\nEMPLOYERS ({cv.Employment.Count}):");
Log(new string('-', 60));
foreach (var emp in cv.Employment)
{
@@ -100,18 +178,18 @@ class Program
var icon = result.IsVerified ? "✓" : "✗";
var period = FormatPeriod(emp.StartDate, emp.EndDate);
Console.WriteLine($"\n {icon} {emp.CompanyName}");
Console.WriteLine($" Period: {period}");
Console.WriteLine($" Role: {emp.JobTitle}");
Log($"\n {icon} {emp.CompanyName}");
Log($" Period: {period}");
Log($" Role: {emp.JobTitle}");
if (result.IsVerified)
{
verifiedEmployers++;
Console.WriteLine($" Match: {result.MatchedCompanyName} ({result.MatchScore}%)");
Log($" Match: {result.MatchedCompanyName} ({result.MatchScore}%)");
if (!string.IsNullOrEmpty(result.MatchedCompanyNumber))
Console.WriteLine($" Company #: {result.MatchedCompanyNumber}");
Log($" Company #: {result.MatchedCompanyNumber}");
if (!string.IsNullOrEmpty(result.CompanyStatus))
Console.WriteLine($" Status: {result.CompanyStatus}");
Log($" Status: {result.CompanyStatus}");
}
else
{
@@ -119,12 +197,12 @@ class Program
}
if (!string.IsNullOrEmpty(result.VerificationNotes))
Console.WriteLine($" Note: {result.VerificationNotes}");
Log($" Note: {result.VerificationNotes}");
}
catch (Exception ex)
{
Console.WriteLine($"\n ✗ {emp.CompanyName}");
Console.WriteLine($" ERROR: {ex.Message}");
Log($"\n ✗ {emp.CompanyName}");
Log($" ERROR: {ex.Message}");
allUnverifiedEmployers.Add(emp.CompanyName);
}
}
@@ -133,8 +211,8 @@ class Program
// Verify Education
if (cv.Education?.Count > 0)
{
Console.WriteLine($"\nEDUCATION ({cv.Education.Count}):");
Console.WriteLine(new string('-', 60));
Log($"\nEDUCATION ({cv.Education.Count}):");
Log(new string('-', 60));
var eduEntries = cv.Education.Select(e => new EducationEntry
{
@@ -152,10 +230,10 @@ class Program
totalEducation++;
var icon = result.IsVerified ? "✓" : "✗";
Console.WriteLine($"\n {icon} {result.ClaimedInstitution}");
Console.WriteLine($" Qualification: {result.ClaimedQualification}");
Log($"\n {icon} {result.ClaimedInstitution}");
Log($" Qualification: {result.ClaimedQualification}");
if (!string.IsNullOrEmpty(result.ClaimedSubject))
Console.WriteLine($" Subject: {result.ClaimedSubject}");
Log($" Subject: {result.ClaimedSubject}");
if (result.IsVerified)
{
@@ -163,41 +241,41 @@ class Program
if (result.MatchedInstitution != null &&
!result.MatchedInstitution.Equals(result.ClaimedInstitution, StringComparison.OrdinalIgnoreCase))
{
Console.WriteLine($" Match: {result.MatchedInstitution}");
Log($" Match: {result.MatchedInstitution}");
}
}
else
{
allUnverifiedInstitutions.Add(result.ClaimedInstitution ?? "Unknown");
Console.WriteLine($" Status: {result.Status}");
Log($" Status: {result.Status}");
}
if (!string.IsNullOrEmpty(result.VerificationNotes))
Console.WriteLine($" Note: {result.VerificationNotes}");
Log($" Note: {result.VerificationNotes}");
}
}
}
catch (Exception ex)
{
errorCount++;
Console.WriteLine($"ERROR processing file: {ex.Message}");
Log($"ERROR processing file: {ex.Message}");
}
}
// Print Summary
Console.WriteLine($"\n\n{new string('=', 80)}");
Console.WriteLine("VERIFICATION SUMMARY");
Console.WriteLine(new string('=', 80));
Log($"\n\n{new string('=', 80)}");
Log("VERIFICATION SUMMARY");
Log(new string('=', 80));
Console.WriteLine($"\nCVs Processed: {processedCount - errorCount}/{cvFiles.Count}");
Log($"\nCVs Processed: {processedCount - errorCount}/{cvFiles.Count}");
if (errorCount > 0)
Console.WriteLine($"Errors: {errorCount}");
Log($"Errors: {errorCount}");
var empRate = totalEmployers > 0 ? verifiedEmployers * 100 / totalEmployers : 0;
var eduRate = totalEducation > 0 ? verifiedEducation * 100 / totalEducation : 0;
Console.WriteLine($"\nEmployers: {verifiedEmployers}/{totalEmployers} verified ({empRate}%)");
Console.WriteLine($"Education: {verifiedEducation}/{totalEducation} verified ({eduRate}%)");
Log($"\nEmployers: {verifiedEmployers}/{totalEmployers} verified ({empRate}%)");
Log($"Education: {verifiedEducation}/{totalEducation} verified ({eduRate}%)");
// List unverified employers
var uniqueUnverifiedEmployers = allUnverifiedEmployers
@@ -208,12 +286,12 @@ class Program
if (uniqueUnverifiedEmployers.Count > 0)
{
Console.WriteLine($"\n{new string('-', 60)}");
Console.WriteLine($"UNVERIFIED EMPLOYERS ({uniqueUnverifiedEmployers.Count} unique):");
Log($"\n{new string('-', 60)}");
Log($"UNVERIFIED EMPLOYERS ({uniqueUnverifiedEmployers.Count} unique):");
foreach (var group in uniqueUnverifiedEmployers)
{
var count = group.Count() > 1 ? $" (x{group.Count()})" : "";
Console.WriteLine($" - {group.Key}{count}");
Log($" - {group.Key}{count}");
}
}
@@ -226,19 +304,30 @@ class Program
if (uniqueUnverifiedInstitutions.Count > 0)
{
Console.WriteLine($"\n{new string('-', 60)}");
Console.WriteLine($"UNVERIFIED INSTITUTIONS ({uniqueUnverifiedInstitutions.Count} unique):");
Log($"\n{new string('-', 60)}");
Log($"UNVERIFIED INSTITUTIONS ({uniqueUnverifiedInstitutions.Count} unique):");
foreach (var group in uniqueUnverifiedInstitutions)
{
var count = group.Count() > 1 ? $" (x{group.Count()})" : "";
Console.WriteLine($" - {group.Key}{count}");
Log($" - {group.Key}{count}");
}
}
Console.WriteLine($"\n{new string('=', 80)}");
Log($"\nCompleted: {DateTime.Now:yyyy-MM-dd HH:mm:ss}");
Log($"\n{new string('=', 80)}");
_logWriter?.Close();
Console.WriteLine($"\nResults written to: {logPath}");
return 0;
}
static void Log(string message)
{
Console.WriteLine(message);
_logWriter?.WriteLine(message);
}
static string AskForFolder()
{
Console.Write("Enter CV folder path: ");
@@ -252,6 +341,57 @@ class Program
return $"{startStr} - {endStr}";
}
static CVData ConvertTestCVData(TestCVData testCv)
{
return new CVData
{
FullName = testCv.Personal?.Name ?? "Unknown",
Email = testCv.Personal?.Email,
Phone = testCv.Personal?.Phone,
Employment = testCv.Employment?.Select(e => new EmploymentEntry
{
CompanyName = e.Company ?? "Unknown",
JobTitle = e.JobTitle ?? "Unknown",
Location = e.Location,
StartDate = ParseDate(e.StartDate),
EndDate = ParseDate(e.EndDate),
IsCurrent = e.EndDate == null,
Description = e.Description
}).ToList() ?? [],
Education = testCv.Education?.Select(e => new EducationEntry
{
Institution = e.Institution ?? "Unknown",
Qualification = e.Qualification,
Subject = e.Subject,
StartDate = ParseDate(e.StartDate),
EndDate = ParseDate(e.EndDate)
}).ToList() ?? [],
Skills = testCv.Skills ?? []
};
}
static DateOnly? ParseDate(string? dateStr)
{
if (string.IsNullOrEmpty(dateStr)) return null;
// Try parsing YYYY-MM format
if (dateStr.Length == 7 && dateStr[4] == '-')
{
if (int.TryParse(dateStr[..4], out var year) && int.TryParse(dateStr[5..], out var month))
{
return new DateOnly(year, month, 1);
}
}
// Try standard parsing
if (DateOnly.TryParse(dateStr, out var date))
{
return date;
}
return null;
}
static void ConfigureServices(IServiceCollection services)
{
// Load configuration - try multiple locations
@@ -263,7 +403,7 @@ class Program
};
var webProjectPath = configPaths.FirstOrDefault(Directory.Exists) ?? "/git/RealCV/src/RealCV.Web";
Console.WriteLine($"Loading config from: {webProjectPath}");
Log($"Loading config from: {webProjectPath}");
var configuration = new ConfigurationBuilder()
.SetBasePath(webProjectPath)
@@ -272,11 +412,14 @@ class Program
.AddJsonFile("appsettings.Production.json", optional: true)
.Build();
// Logging - minimal output
// Logging - show info level for verification details
services.AddLogging(builder =>
{
builder.AddConsole();
builder.SetMinimumLevel(LogLevel.Warning);
builder.SetMinimumLevel(LogLevel.Information);
// Filter out noisy libraries
builder.AddFilter("Microsoft", LogLevel.Warning);
builder.AddFilter("System", LogLevel.Warning);
});
// Database