feat: Add AI-powered compound company name splitting

Uses Claude Haiku to intelligently detect when a company name contains
multiple companies (e.g., "ASDA/WALMART", "Corus & Laura Ashley Hotels")
vs single companies with similar patterns (e.g., "Ernst & Young").

- Adds ExtractCompanyNamesAsync to ICompanyNameMatcherService
- Only triggers for names with potential separators (/, &, "and")
- Verifies each extracted part individually, returns first match
- Uses fast Haiku model to minimize cost

Results:
- ASDA/WALMART → verified via 'ASDA' → ASDA GROUP LIMITED
- Corus & Laura Ashley Hotels → verified via 'Corus' → Tata Steel UK
- Employers: 104/120 verified (86%)

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

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
2026-01-22 11:04:30 +00:00
parent 27921d625f
commit 94ca6e1b9a
7 changed files with 1025 additions and 0 deletions

View File

@@ -12,4 +12,13 @@ public interface ICompanyNameMatcherService
string cvCompanyName,
List<CompanyCandidate> candidates,
CancellationToken cancellationToken = default);
/// <summary>
/// Uses AI to detect if a company name contains multiple companies and extract them.
/// Returns null or single-item list if it's a single company (e.g., "Ernst & Young").
/// Returns multiple items if compound (e.g., "ASDA/WALMART" -> ["ASDA", "WALMART"]).
/// </summary>
Task<List<string>?> ExtractCompanyNamesAsync(
string companyName,
CancellationToken cancellationToken = default);
}

View File

@@ -207,4 +207,107 @@ public sealed class AICompanyNameMatcherService : ICompanyNameMatcherService
return null; // Fall back to fuzzy matching
}
}
private const string CompoundNamePrompt = """
Analyze this company name from a CV and determine if it refers to ONE company or MULTIPLE companies.
Company name: "{COMPANY_NAME}"
Examples:
- "Ernst & Young" ONE company (it's the full name of the accounting firm)
- "Marks & Spencer" ONE company (it's the full name of the retailer)
- "ASDA/WALMART" TWO companies: ["ASDA", "WALMART"] (person worked at both or it's showing ownership)
- "Corus & Laura Ashley Hotels" TWO companies: ["Corus", "Laura Ashley Hotels"] (different industries)
- "PwC" ONE company
- "Deloitte and Touche" ONE company (historical name of Deloitte)
- "BMW Group Ireland" ONE company
- "Tesco Stores and Distribution" ONE company (departments of same company)
Rules:
1. Well-known company names with "&" or "and" are SINGLE companies (Ernst & Young, Marks & Spencer, Procter & Gamble)
2. A "/" usually indicates multiple companies or ownership relationship
3. If the parts are in completely different industries, they're likely separate companies
4. If one part is clearly a subsidiary/department of the other, treat as ONE company
Respond with ONLY valid JSON:
{
"isSingleCompany": boolean,
"companies": ["company1", "company2"] or ["single company name"],
"reasoning": "brief explanation"
}
""";
public async Task<List<string>?> ExtractCompanyNamesAsync(
string companyName,
CancellationToken cancellationToken = default)
{
if (string.IsNullOrWhiteSpace(companyName))
{
return null;
}
_logger.LogDebug("Using AI to check if '{CompanyName}' is a compound name", companyName);
try
{
var prompt = CompoundNamePrompt.Replace("{COMPANY_NAME}", companyName);
var messages = new List<Message>
{
new(RoleType.User, prompt)
};
var parameters = new MessageParameters
{
Model = "claude-3-5-haiku-20241022",
MaxTokens = 256,
Messages = messages,
System = [new SystemMessage("You are a company name parser. Respond only with valid JSON.")]
};
var response = await _anthropicClient.Messages.GetClaudeMessageAsync(parameters, cancellationToken);
var responseText = response.Content
.OfType<TextContent>()
.FirstOrDefault()?.Text;
if (string.IsNullOrWhiteSpace(responseText))
{
_logger.LogWarning("AI returned empty response for compound name check");
return null;
}
responseText = JsonResponseHelper.CleanJsonResponse(responseText);
var result = JsonSerializer.Deserialize<CompoundNameResponse>(responseText, JsonDefaults.CamelCase);
if (result is null)
{
_logger.LogWarning("Failed to deserialize compound name response: {Response}", responseText);
return null;
}
_logger.LogDebug("AI compound name result: IsSingle={IsSingle}, Companies=[{Companies}], Reasoning={Reasoning}",
result.IsSingleCompany, string.Join(", ", result.Companies ?? []), result.Reasoning);
if (result.IsSingleCompany || result.Companies is null || result.Companies.Count < 2)
{
return null; // Single company, no splitting needed
}
return result.Companies;
}
catch (Exception ex)
{
_logger.LogError(ex, "AI compound name detection failed for '{CompanyName}'", companyName);
return null;
}
}
private sealed class CompoundNameResponse
{
public bool IsSingleCompany { get; set; }
public List<string>? Companies { get; set; }
public string? Reasoning { get; set; }
}
}

View File

@@ -226,6 +226,14 @@ public sealed class CompanyVerifierService : ICompanyVerifierService
};
}
// Check 4: Is this a compound company name (e.g., "ASDA/WALMART", "Corus & Laura Ashley Hotels")?
// Try to verify each part individually
var compoundResult = await TryVerifyCompoundNameAsync(normalizedName, companyName, startDate, endDate, jobTitle, flags);
if (compoundResult is not null)
{
return compoundResult;
}
// Try to find a cached match first (but only if it existed at claimed start date)
var cachedMatch = await FindCachedMatchAsync(normalizedName);
if (cachedMatch is not null)
@@ -833,6 +841,70 @@ public sealed class CompanyVerifierService : ICompanyVerifierService
return normalized;
}
/// <summary>
/// Attempts to verify compound company names by detecting if multiple companies are mentioned.
/// Only triggers for names with potential separators (/, &amp;, "and") to avoid unnecessary AI calls.
/// </summary>
private async Task<CompanyVerificationResult?> TryVerifyCompoundNameAsync(
string normalizedName,
string originalName,
DateOnly? startDate,
DateOnly? endDate,
string? jobTitle,
List<CompanyVerificationFlag> flags)
{
// Quick check: only process names that might be compound
// Look for separators that could indicate multiple companies
var hasPotentialSeparator = normalizedName.Contains('/')
|| normalizedName.Contains(" & ")
|| normalizedName.Contains(" and ", StringComparison.OrdinalIgnoreCase);
if (!hasPotentialSeparator)
{
return null;
}
// Use AI to determine if this is a compound name and extract parts
var extractedParts = await _aiMatcher.ExtractCompanyNamesAsync(normalizedName);
if (extractedParts is null || extractedParts.Count < 2)
{
// AI determined this is a single company (e.g., "Ernst & Young")
return null;
}
_logger.LogDebug("AI detected compound company name '{Name}', extracted parts: {Parts}",
originalName, string.Join(", ", extractedParts.Select(p => $"'{p}'")));
// Try to verify each extracted part - return success on first match
foreach (var part in extractedParts)
{
// Skip parts that are too short
if (part.Length < 3) continue;
_logger.LogDebug("Trying to verify compound part: '{Part}'", part);
// Recursively verify this part
var partResult = await VerifyCompanyAsync(part, startDate, endDate, jobTitle);
if (partResult.IsVerified)
{
_logger.LogInformation("Compound name '{Original}' verified via part '{Part}' -> {Match}",
originalName, part, partResult.MatchedCompanyName);
return partResult with
{
ClaimedCompany = originalName,
VerificationNotes = $"Verified via '{part}': {partResult.VerificationNotes ?? partResult.MatchedCompanyName}"
};
}
}
// None of the parts could be verified
_logger.LogDebug("No parts of compound name '{Name}' could be verified", originalName);
return null;
}
private async Task<CompanyCache?> FindCachedMatchAsync(string companyName)
{
var cutoffDate = DateTime.UtcNow.AddDays(-CacheExpirationDays);

View File

@@ -0,0 +1,329 @@
using System.Text.Json;
using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.Configuration;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Logging;
using RealCV.Application.Interfaces;
using RealCV.Application.Models;
using RealCV.Infrastructure.Data;
using RealCV.Infrastructure.ExternalApis;
using RealCV.Infrastructure.Services;
using RealCV.Infrastructure.Configuration;
namespace RealCV.Tests.Integration;
/// <summary>
/// Test utility to batch process CVs and output verification findings.
/// Run with: dotnet test --filter "FullyQualifiedName~CVBatchTester" -- TestRunParameters.Parameter(name=\"CvFolder\", value=\"/path/to/cvs\")
/// Or use the ProcessFolder method directly.
/// </summary>
public class CVBatchTester
{
private readonly IServiceProvider _serviceProvider;
public CVBatchTester()
{
var services = new ServiceCollection();
ConfigureServices(services);
_serviceProvider = services.BuildServiceProvider();
}
private static void ConfigureServices(IServiceCollection services)
{
// Load configuration
var configuration = new ConfigurationBuilder()
.SetBasePath(Directory.GetCurrentDirectory())
.AddJsonFile("appsettings.json", optional: true)
.AddJsonFile("appsettings.Development.json", optional: true)
.AddEnvironmentVariables()
.Build();
// Logging
services.AddLogging(builder =>
{
builder.AddConsole();
builder.SetMinimumLevel(LogLevel.Information);
});
// Database
var connectionString = configuration.GetConnectionString("DefaultConnection")
?? "Server=127.0.0.1;Database=RealCV;User Id=SA;Password=TrueCV_Sql2024!;TrustServerCertificate=True";
services.AddDbContextFactory<ApplicationDbContext>(options =>
options.UseSqlServer(connectionString));
// Companies House
services.Configure<CompaniesHouseSettings>(options =>
{
options.BaseUrl = configuration["CompaniesHouse:BaseUrl"] ?? "https://api.company-information.service.gov.uk";
options.ApiKey = configuration["CompaniesHouse:ApiKey"] ?? "";
});
services.AddHttpClient<CompaniesHouseClient>();
// Anthropic (for AI matching)
services.Configure<AnthropicSettings>(options =>
{
options.ApiKey = configuration["Anthropic:ApiKey"] ?? "";
});
services.AddHttpClient<AnthropicClient>();
services.AddScoped<ICompanyNameMatcherService, CompanyNameMatcherService>();
// Services
services.AddScoped<ICompanyVerifierService, CompanyVerifierService>();
services.AddScoped<IEducationVerifierService, EducationVerifierService>();
services.AddScoped<ICVParserService, CVParserService>();
}
/// <summary>
/// Process all CVs in a folder and return verification results.
/// </summary>
public async Task<List<CVVerificationSummary>> ProcessFolderAsync(string folderPath)
{
if (!Directory.Exists(folderPath))
{
throw new DirectoryNotFoundException($"Folder not found: {folderPath}");
}
var cvFiles = Directory.GetFiles(folderPath, "*.*", SearchOption.TopDirectoryOnly)
.Where(f => f.EndsWith(".pdf", StringComparison.OrdinalIgnoreCase) ||
f.EndsWith(".docx", StringComparison.OrdinalIgnoreCase) ||
f.EndsWith(".doc", StringComparison.OrdinalIgnoreCase))
.ToList();
Console.WriteLine($"Found {cvFiles.Count} CV files in {folderPath}");
Console.WriteLine(new string('=', 80));
var results = new List<CVVerificationSummary>();
foreach (var cvFile in cvFiles)
{
Console.WriteLine($"\nProcessing: {Path.GetFileName(cvFile)}");
Console.WriteLine(new string('-', 60));
try
{
var result = await ProcessSingleCVAsync(cvFile);
results.Add(result);
PrintSummary(result);
}
catch (Exception ex)
{
Console.WriteLine($"ERROR: {ex.Message}");
results.Add(new CVVerificationSummary
{
FileName = Path.GetFileName(cvFile),
Error = ex.Message
});
}
}
// Print overall summary
Console.WriteLine("\n" + new string('=', 80));
Console.WriteLine("OVERALL SUMMARY");
Console.WriteLine(new string('=', 80));
PrintOverallSummary(results);
return results;
}
private async Task<CVVerificationSummary> ProcessSingleCVAsync(string filePath)
{
using var scope = _serviceProvider.CreateScope();
var cvParser = scope.ServiceProvider.GetRequiredService<ICVParserService>();
var companyVerifier = scope.ServiceProvider.GetRequiredService<ICompanyVerifierService>();
var educationVerifier = scope.ServiceProvider.GetRequiredService<IEducationVerifierService>();
// Parse the CV
await using var fileStream = File.OpenRead(filePath);
var parsedCV = await cvParser.ParseAsync(fileStream, Path.GetFileName(filePath));
var summary = new CVVerificationSummary
{
FileName = Path.GetFileName(filePath),
CandidateName = parsedCV.PersonalInfo?.FullName ?? "Unknown"
};
// Verify employers
if (parsedCV.Employment?.Count > 0)
{
foreach (var employment in parsedCV.Employment)
{
try
{
var result = await companyVerifier.VerifyCompanyAsync(
employment.CompanyName,
employment.StartDate,
employment.EndDate,
employment.JobTitle);
summary.EmployerResults.Add(new EmployerVerificationSummary
{
ClaimedName = employment.CompanyName,
MatchedName = result.MatchedCompanyName,
CompanyNumber = result.MatchedCompanyNumber,
IsVerified = result.IsVerified,
MatchScore = result.MatchScore,
Notes = result.VerificationNotes,
Status = result.CompanyStatus
});
}
catch (Exception ex)
{
summary.EmployerResults.Add(new EmployerVerificationSummary
{
ClaimedName = employment.CompanyName,
IsVerified = false,
Notes = $"Error: {ex.Message}"
});
}
}
}
// Verify education
if (parsedCV.Education?.Count > 0)
{
var educationResults = educationVerifier.VerifyAll(
parsedCV.Education.Select(e => new EducationEntry
{
Institution = e.Institution,
Qualification = e.Qualification,
Subject = e.Subject,
StartDate = e.StartDate,
EndDate = e.EndDate
}).ToList());
foreach (var result in educationResults)
{
summary.EducationResults.Add(new EducationVerificationSummary
{
ClaimedInstitution = result.ClaimedInstitution,
MatchedInstitution = result.MatchedInstitution,
Qualification = result.ClaimedQualification,
IsVerified = result.IsVerified,
Status = result.Status,
Notes = result.VerificationNotes
});
}
}
return summary;
}
private static void PrintSummary(CVVerificationSummary summary)
{
Console.WriteLine($"Candidate: {summary.CandidateName}");
Console.WriteLine($"\n EMPLOYERS ({summary.EmployerResults.Count}):");
foreach (var emp in summary.EmployerResults)
{
var status = emp.IsVerified ? "✓" : "✗";
var matchInfo = emp.IsVerified
? $"-> {emp.MatchedName} ({emp.MatchScore}%)"
: emp.Notes ?? "Not found";
Console.WriteLine($" {status} {emp.ClaimedName}");
Console.WriteLine($" {matchInfo}");
}
Console.WriteLine($"\n EDUCATION ({summary.EducationResults.Count}):");
foreach (var edu in summary.EducationResults)
{
var status = edu.IsVerified ? "✓" : "✗";
var matchInfo = edu.IsVerified && edu.MatchedInstitution != null
? $"-> {edu.MatchedInstitution}"
: edu.Notes ?? edu.Status;
Console.WriteLine($" {status} {edu.ClaimedInstitution}");
Console.WriteLine($" {edu.Qualification}");
Console.WriteLine($" {matchInfo}");
}
}
private static void PrintOverallSummary(List<CVVerificationSummary> results)
{
var successfulCVs = results.Count(r => r.Error == null);
var totalEmployers = results.Sum(r => r.EmployerResults.Count);
var verifiedEmployers = results.Sum(r => r.EmployerResults.Count(e => e.IsVerified));
var totalEducation = results.Sum(r => r.EducationResults.Count);
var verifiedEducation = results.Sum(r => r.EducationResults.Count(e => e.IsVerified));
Console.WriteLine($"CVs Processed: {successfulCVs}/{results.Count}");
Console.WriteLine($"Employers: {verifiedEmployers}/{totalEmployers} verified ({(totalEmployers > 0 ? verifiedEmployers * 100 / totalEmployers : 0)}%)");
Console.WriteLine($"Education: {verifiedEducation}/{totalEducation} verified ({(totalEducation > 0 ? verifiedEducation * 100 / totalEducation : 0)}%)");
// List unverified employers
var unverifiedEmployers = results
.SelectMany(r => r.EmployerResults.Where(e => !e.IsVerified))
.GroupBy(e => e.ClaimedName)
.OrderByDescending(g => g.Count())
.ToList();
if (unverifiedEmployers.Count > 0)
{
Console.WriteLine($"\nUNVERIFIED EMPLOYERS ({unverifiedEmployers.Count} unique):");
foreach (var group in unverifiedEmployers.Take(20))
{
Console.WriteLine($" - {group.Key} (x{group.Count()})");
}
}
// List unverified institutions
var unverifiedEducation = results
.SelectMany(r => r.EducationResults.Where(e => !e.IsVerified))
.GroupBy(e => e.ClaimedInstitution)
.OrderByDescending(g => g.Count())
.ToList();
if (unverifiedEducation.Count > 0)
{
Console.WriteLine($"\nUNVERIFIED INSTITUTIONS ({unverifiedEducation.Count} unique):");
foreach (var group in unverifiedEducation.Take(20))
{
Console.WriteLine($" - {group.Key} (x{group.Count()})");
}
}
}
/// <summary>
/// Export results to JSON for further analysis.
/// </summary>
public static void ExportToJson(List<CVVerificationSummary> results, string outputPath)
{
var json = JsonSerializer.Serialize(results, new JsonSerializerOptions
{
WriteIndented = true
});
File.WriteAllText(outputPath, json);
Console.WriteLine($"\nResults exported to: {outputPath}");
}
}
public class CVVerificationSummary
{
public string FileName { get; set; } = "";
public string CandidateName { get; set; } = "";
public string? Error { get; set; }
public List<EmployerVerificationSummary> EmployerResults { get; set; } = new();
public List<EducationVerificationSummary> EducationResults { get; set; } = new();
}
public class EmployerVerificationSummary
{
public string ClaimedName { get; set; } = "";
public string? MatchedName { get; set; }
public string? CompanyNumber { get; set; }
public bool IsVerified { get; set; }
public int MatchScore { get; set; }
public string? Notes { get; set; }
public string? Status { get; set; }
}
public class EducationVerificationSummary
{
public string ClaimedInstitution { get; set; } = "";
public string? MatchedInstitution { get; set; }
public string? Qualification { get; set; }
public bool IsVerified { get; set; }
public string? Status { get; set; }
public string? Notes { get; set; }
}

View File

@@ -0,0 +1,15 @@
<Project Sdk="Microsoft.NET.Sdk">
<PropertyGroup>
<OutputType>Exe</OutputType>
<TargetFramework>net8.0</TargetFramework>
<ImplicitUsings>enable</ImplicitUsings>
<Nullable>enable</Nullable>
</PropertyGroup>
<ItemGroup>
<ProjectReference Include="../../src/RealCV.Application/RealCV.Application.csproj" />
<ProjectReference Include="../../src/RealCV.Infrastructure/RealCV.Infrastructure.csproj" />
</ItemGroup>
</Project>

View File

@@ -0,0 +1,302 @@
using System.Text.Json;
using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.Configuration;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Logging;
using RealCV.Application.Interfaces;
using RealCV.Application.Models;
using RealCV.Infrastructure.Configuration;
using RealCV.Infrastructure.Data;
using RealCV.Infrastructure.ExternalApis;
using RealCV.Infrastructure.Services;
namespace CVBatchTester;
class Program
{
static async Task<int> Main(string[] args)
{
var folderPath = args.FirstOrDefault() ?? AskForFolder();
if (string.IsNullOrEmpty(folderPath) || !Directory.Exists(folderPath))
{
Console.WriteLine($"Error: Folder not found: {folderPath}");
Console.WriteLine("Usage: CVBatchTester <folder-path>");
Console.WriteLine(" e.g. CVBatchTester /home/user/cvs");
return 1;
}
Console.WriteLine($"CV Batch Verification Tester");
Console.WriteLine($"Processing CVs from: {folderPath}");
Console.WriteLine(new string('=', 80));
// Setup DI
var services = new ServiceCollection();
ConfigureServices(services);
var provider = services.BuildServiceProvider();
// Find CV files
var cvFiles = Directory.GetFiles(folderPath, "*.*", SearchOption.TopDirectoryOnly)
.Where(f => f.EndsWith(".pdf", StringComparison.OrdinalIgnoreCase) ||
f.EndsWith(".docx", StringComparison.OrdinalIgnoreCase) ||
f.EndsWith(".doc", StringComparison.OrdinalIgnoreCase))
.OrderBy(f => f)
.ToList();
Console.WriteLine($"Found {cvFiles.Count} CV files\n");
if (cvFiles.Count == 0)
{
Console.WriteLine("No CV files found (.pdf, .docx, .doc)");
return 1;
}
// Track results
var allUnverifiedEmployers = new List<string>();
var allUnverifiedInstitutions = new List<string>();
var totalEmployers = 0;
var verifiedEmployers = 0;
var totalEducation = 0;
var verifiedEducation = 0;
var processedCount = 0;
var errorCount = 0;
foreach (var cvFile in cvFiles)
{
Console.WriteLine($"\n{new string('=', 80)}");
Console.WriteLine($"[{++processedCount}/{cvFiles.Count}] {Path.GetFileName(cvFile)}");
Console.WriteLine(new string('=', 80));
try
{
using var scope = provider.CreateScope();
var parser = scope.ServiceProvider.GetRequiredService<ICVParserService>();
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));
Console.WriteLine($"Candidate: {cv.FullName}");
// Verify Employers
if (cv.Employment?.Count > 0)
{
Console.WriteLine($"\nEMPLOYERS ({cv.Employment.Count}):");
Console.WriteLine(new string('-', 60));
foreach (var emp in cv.Employment)
{
totalEmployers++;
try
{
var result = await companyVerifier.VerifyCompanyAsync(
emp.CompanyName,
emp.StartDate,
emp.EndDate,
emp.JobTitle);
var icon = result.IsVerified ? "✓" : "✗";
var period = FormatPeriod(emp.StartDate, emp.EndDate);
Console.WriteLine($"\n {icon} {emp.CompanyName}");
Console.WriteLine($" Period: {period}");
Console.WriteLine($" Role: {emp.JobTitle}");
if (result.IsVerified)
{
verifiedEmployers++;
Console.WriteLine($" Match: {result.MatchedCompanyName} ({result.MatchScore}%)");
if (!string.IsNullOrEmpty(result.MatchedCompanyNumber))
Console.WriteLine($" Company #: {result.MatchedCompanyNumber}");
if (!string.IsNullOrEmpty(result.CompanyStatus))
Console.WriteLine($" Status: {result.CompanyStatus}");
}
else
{
allUnverifiedEmployers.Add(emp.CompanyName);
}
if (!string.IsNullOrEmpty(result.VerificationNotes))
Console.WriteLine($" Note: {result.VerificationNotes}");
}
catch (Exception ex)
{
Console.WriteLine($"\n ✗ {emp.CompanyName}");
Console.WriteLine($" ERROR: {ex.Message}");
allUnverifiedEmployers.Add(emp.CompanyName);
}
}
}
// Verify Education
if (cv.Education?.Count > 0)
{
Console.WriteLine($"\nEDUCATION ({cv.Education.Count}):");
Console.WriteLine(new string('-', 60));
var eduEntries = cv.Education.Select(e => new EducationEntry
{
Institution = e.Institution,
Qualification = e.Qualification,
Subject = e.Subject,
StartDate = e.StartDate,
EndDate = e.EndDate
}).ToList();
var eduResults = eduVerifier.VerifyAll(eduEntries);
foreach (var result in eduResults)
{
totalEducation++;
var icon = result.IsVerified ? "✓" : "✗";
Console.WriteLine($"\n {icon} {result.ClaimedInstitution}");
Console.WriteLine($" Qualification: {result.ClaimedQualification}");
if (!string.IsNullOrEmpty(result.ClaimedSubject))
Console.WriteLine($" Subject: {result.ClaimedSubject}");
if (result.IsVerified)
{
verifiedEducation++;
if (result.MatchedInstitution != null &&
!result.MatchedInstitution.Equals(result.ClaimedInstitution, StringComparison.OrdinalIgnoreCase))
{
Console.WriteLine($" Match: {result.MatchedInstitution}");
}
}
else
{
allUnverifiedInstitutions.Add(result.ClaimedInstitution ?? "Unknown");
Console.WriteLine($" Status: {result.Status}");
}
if (!string.IsNullOrEmpty(result.VerificationNotes))
Console.WriteLine($" Note: {result.VerificationNotes}");
}
}
}
catch (Exception ex)
{
errorCount++;
Console.WriteLine($"ERROR processing file: {ex.Message}");
}
}
// Print Summary
Console.WriteLine($"\n\n{new string('=', 80)}");
Console.WriteLine("VERIFICATION SUMMARY");
Console.WriteLine(new string('=', 80));
Console.WriteLine($"\nCVs Processed: {processedCount - errorCount}/{cvFiles.Count}");
if (errorCount > 0)
Console.WriteLine($"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}%)");
// List unverified employers
var uniqueUnverifiedEmployers = allUnverifiedEmployers
.GroupBy(e => e, StringComparer.OrdinalIgnoreCase)
.OrderByDescending(g => g.Count())
.ThenBy(g => g.Key)
.ToList();
if (uniqueUnverifiedEmployers.Count > 0)
{
Console.WriteLine($"\n{new string('-', 60)}");
Console.WriteLine($"UNVERIFIED EMPLOYERS ({uniqueUnverifiedEmployers.Count} unique):");
foreach (var group in uniqueUnverifiedEmployers)
{
var count = group.Count() > 1 ? $" (x{group.Count()})" : "";
Console.WriteLine($" - {group.Key}{count}");
}
}
// List unverified institutions
var uniqueUnverifiedInstitutions = allUnverifiedInstitutions
.GroupBy(i => i, StringComparer.OrdinalIgnoreCase)
.OrderByDescending(g => g.Count())
.ThenBy(g => g.Key)
.ToList();
if (uniqueUnverifiedInstitutions.Count > 0)
{
Console.WriteLine($"\n{new string('-', 60)}");
Console.WriteLine($"UNVERIFIED INSTITUTIONS ({uniqueUnverifiedInstitutions.Count} unique):");
foreach (var group in uniqueUnverifiedInstitutions)
{
var count = group.Count() > 1 ? $" (x{group.Count()})" : "";
Console.WriteLine($" - {group.Key}{count}");
}
}
Console.WriteLine($"\n{new string('=', 80)}");
return 0;
}
static string AskForFolder()
{
Console.Write("Enter CV folder path: ");
return Console.ReadLine() ?? "";
}
static string FormatPeriod(DateOnly? start, DateOnly? end)
{
var startStr = start?.ToString("MMM yyyy") ?? "?";
var endStr = end?.ToString("MMM yyyy") ?? "Present";
return $"{startStr} - {endStr}";
}
static void ConfigureServices(IServiceCollection services)
{
// Load configuration - try multiple locations
var configPaths = new[]
{
"/var/www/realcv",
"/git/RealCV/src/RealCV.Web",
Path.GetFullPath(Path.Combine(AppContext.BaseDirectory, "..", "..", "..", "..", "..", "src", "RealCV.Web"))
};
var webProjectPath = configPaths.FirstOrDefault(Directory.Exists) ?? "/git/RealCV/src/RealCV.Web";
Console.WriteLine($"Loading config from: {webProjectPath}");
var configuration = new ConfigurationBuilder()
.SetBasePath(webProjectPath)
.AddJsonFile("appsettings.json", optional: true)
.AddJsonFile("appsettings.Development.json", optional: true)
.AddJsonFile("appsettings.Production.json", optional: true)
.Build();
// Logging - minimal output
services.AddLogging(builder =>
{
builder.AddConsole();
builder.SetMinimumLevel(LogLevel.Warning);
});
// Database
var connectionString = configuration.GetConnectionString("DefaultConnection")
?? "Server=127.0.0.1;Database=RealCV;User Id=SA;Password=TrueCV_Sql2024!;TrustServerCertificate=True";
services.AddDbContextFactory<ApplicationDbContext>(options =>
options.UseSqlServer(connectionString));
// Companies House - use configuration binding
services.Configure<CompaniesHouseSettings>(configuration.GetSection(CompaniesHouseSettings.SectionName));
services.AddHttpClient<CompaniesHouseClient>();
// Anthropic - use configuration binding
services.Configure<AnthropicSettings>(configuration.GetSection(AnthropicSettings.SectionName));
services.AddScoped<ICompanyNameMatcherService, AICompanyNameMatcherService>();
// Services
services.AddScoped<ICompanyVerifierService, CompanyVerifierService>();
services.AddScoped<IEducationVerifierService, EducationVerifierService>();
services.AddScoped<ICVParserService, CVParserService>();
}
}

195
tools/batch-test-cvs.cs Normal file
View File

@@ -0,0 +1,195 @@
#!/usr/bin/env dotnet-script
#r "nuget: Microsoft.EntityFrameworkCore.SqlServer, 8.0.0"
#r "nuget: Microsoft.Extensions.Configuration.Json, 8.0.0"
#r "nuget: Microsoft.Extensions.DependencyInjection, 8.0.0"
#r "nuget: Microsoft.Extensions.Logging.Console, 8.0.0"
#r "../src/RealCV.Application/bin/Debug/net8.0/RealCV.Application.dll"
#r "../src/RealCV.Infrastructure/bin/Debug/net8.0/RealCV.Infrastructure.dll"
#r "../src/RealCV.Domain/bin/Debug/net8.0/RealCV.Domain.dll"
// This is a dotnet-script file. Run with: dotnet script batch-test-cvs.cs -- /path/to/cvs
// Install dotnet-script: dotnet tool install -g dotnet-script
using System;
using System.IO;
using System.Linq;
using System.Threading.Tasks;
using System.Collections.Generic;
using System.Text.Json;
using Microsoft.EntityFrameworkCore;
using Microsoft.Extensions.Configuration;
using Microsoft.Extensions.DependencyInjection;
using Microsoft.Extensions.Logging;
using RealCV.Application.Interfaces;
using RealCV.Application.Models;
using RealCV.Infrastructure.Data;
using RealCV.Infrastructure.Services;
using RealCV.Infrastructure.ExternalApis;
using RealCV.Infrastructure.Configuration;
var folderPath = Args.FirstOrDefault() ?? "/tmp/test-cvs";
if (!Directory.Exists(folderPath))
{
Console.WriteLine($"Error: Folder not found: {folderPath}");
Console.WriteLine("Usage: dotnet script batch-test-cvs.cs -- /path/to/cvs");
return 1;
}
Console.WriteLine($"Processing CVs from: {folderPath}");
Console.WriteLine(new string('=', 80));
// Setup DI
var services = new ServiceCollection();
var configuration = new ConfigurationBuilder()
.SetBasePath(Path.Combine(Directory.GetCurrentDirectory(), "../src/RealCV.Web"))
.AddJsonFile("appsettings.json", optional: true)
.AddJsonFile("appsettings.Development.json", optional: true)
.Build();
services.AddLogging(b => b.AddConsole().SetMinimumLevel(LogLevel.Warning));
services.AddDbContextFactory<ApplicationDbContext>(options =>
options.UseSqlServer(configuration.GetConnectionString("DefaultConnection")));
services.Configure<CompaniesHouseSettings>(configuration.GetSection("CompaniesHouse"));
services.Configure<AnthropicSettings>(configuration.GetSection("Anthropic"));
services.AddHttpClient<CompaniesHouseClient>();
services.AddHttpClient<AnthropicClient>();
services.AddScoped<ICompanyNameMatcherService, CompanyNameMatcherService>();
services.AddScoped<ICompanyVerifierService, CompanyVerifierService>();
services.AddScoped<IEducationVerifierService, EducationVerifierService>();
services.AddScoped<ICVParserService, CVParserService>();
var provider = services.BuildServiceProvider();
var cvFiles = Directory.GetFiles(folderPath, "*.*")
.Where(f => f.EndsWith(".pdf", StringComparison.OrdinalIgnoreCase) ||
f.EndsWith(".docx", StringComparison.OrdinalIgnoreCase))
.ToList();
Console.WriteLine($"Found {cvFiles.Count} CV files\n");
var allUnverifiedEmployers = new List<string>();
var allUnverifiedInstitutions = new List<string>();
var totalEmployers = 0;
var verifiedEmployers = 0;
var totalEducation = 0;
var verifiedEducation = 0;
foreach (var cvFile in cvFiles)
{
Console.WriteLine($"\n{'=',-80}");
Console.WriteLine($"FILE: {Path.GetFileName(cvFile)}");
Console.WriteLine($"{'=',-80}");
try
{
using var scope = provider.CreateScope();
var parser = scope.ServiceProvider.GetRequiredService<ICVParserService>();
var companyVerifier = scope.ServiceProvider.GetRequiredService<ICompanyVerifierService>();
var eduVerifier = scope.ServiceProvider.GetRequiredService<IEducationVerifierService>();
await using var stream = File.OpenRead(cvFile);
var cv = await parser.ParseAsync(stream, Path.GetFileName(cvFile));
Console.WriteLine($"Candidate: {cv.PersonalInfo?.FullName ?? "Unknown"}");
// Employers
if (cv.Employment?.Count > 0)
{
Console.WriteLine($"\nEMPLOYERS ({cv.Employment.Count}):");
foreach (var emp in cv.Employment)
{
totalEmployers++;
var result = await companyVerifier.VerifyCompanyAsync(
emp.CompanyName, emp.StartDate, emp.EndDate, emp.JobTitle);
var icon = result.IsVerified ? "✓" : "✗";
Console.WriteLine($" {icon} {emp.CompanyName}");
if (result.IsVerified)
{
verifiedEmployers++;
Console.WriteLine($" → {result.MatchedCompanyName} ({result.MatchScore}%)");
if (!string.IsNullOrEmpty(result.VerificationNotes))
Console.WriteLine($" Note: {result.VerificationNotes}");
}
else
{
allUnverifiedEmployers.Add(emp.CompanyName);
Console.WriteLine($" Note: {result.VerificationNotes ?? "Not found"}");
}
}
}
// Education
if (cv.Education?.Count > 0)
{
Console.WriteLine($"\nEDUCATION ({cv.Education.Count}):");
var eduEntries = cv.Education.Select(e => new EducationEntry
{
Institution = e.Institution,
Qualification = e.Qualification,
Subject = e.Subject,
StartDate = e.StartDate,
EndDate = e.EndDate
}).ToList();
var eduResults = eduVerifier.VerifyAll(eduEntries);
foreach (var result in eduResults)
{
totalEducation++;
var icon = result.IsVerified ? "✓" : "✗";
Console.WriteLine($" {icon} {result.ClaimedInstitution}");
Console.WriteLine($" {result.ClaimedQualification}");
if (result.IsVerified)
{
verifiedEducation++;
if (result.MatchedInstitution != null && result.MatchedInstitution != result.ClaimedInstitution)
Console.WriteLine($" → {result.MatchedInstitution}");
}
else
{
allUnverifiedInstitutions.Add(result.ClaimedInstitution ?? "Unknown");
Console.WriteLine($" Status: {result.Status}");
if (!string.IsNullOrEmpty(result.VerificationNotes))
Console.WriteLine($" Note: {result.VerificationNotes}");
}
}
}
}
catch (Exception ex)
{
Console.WriteLine($"ERROR: {ex.Message}");
}
}
// Summary
Console.WriteLine($"\n\n{'=',-80}");
Console.WriteLine("SUMMARY");
Console.WriteLine($"{'=',-80}");
Console.WriteLine($"CVs Processed: {cvFiles.Count}");
Console.WriteLine($"Employers: {verifiedEmployers}/{totalEmployers} verified ({(totalEmployers > 0 ? verifiedEmployers * 100 / totalEmployers : 0)}%)");
Console.WriteLine($"Education: {verifiedEducation}/{totalEducation} verified ({(totalEducation > 0 ? verifiedEducation * 100 / totalEducation : 0)}%)");
var uniqueUnverifiedEmployers = allUnverifiedEmployers.Distinct().OrderBy(x => x).ToList();
if (uniqueUnverifiedEmployers.Count > 0)
{
Console.WriteLine($"\nUNVERIFIED EMPLOYERS ({uniqueUnverifiedEmployers.Count}):");
foreach (var emp in uniqueUnverifiedEmployers)
Console.WriteLine($" - {emp}");
}
var uniqueUnverifiedInstitutions = allUnverifiedInstitutions.Distinct().OrderBy(x => x).ToList();
if (uniqueUnverifiedInstitutions.Count > 0)
{
Console.WriteLine($"\nUNVERIFIED INSTITUTIONS ({uniqueUnverifiedInstitutions.Count}):");
foreach (var inst in uniqueUnverifiedInstitutions)
Console.WriteLine($" - {inst}");
}
return 0;