chore: Remove OpenCorporates integration (requires paid plan for commercial use)
Keep only the genuinely free APIs: - FCA Register (free with registration) - SRA Register (free public API) - GitHub (free tier sufficient) - ORCID (free public API) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
@@ -1,376 +0,0 @@
|
||||
using Microsoft.Extensions.Logging;
|
||||
using RealCV.Application.Interfaces;
|
||||
using RealCV.Application.Models;
|
||||
using RealCV.Infrastructure.Clients;
|
||||
|
||||
namespace RealCV.Infrastructure.Services;
|
||||
|
||||
public sealed class InternationalCompanyVerifierService : IInternationalCompanyVerifierService
|
||||
{
|
||||
private readonly OpenCorporatesClient _openCorporatesClient;
|
||||
private readonly ILogger<InternationalCompanyVerifierService> _logger;
|
||||
|
||||
// Common jurisdiction codes
|
||||
private static readonly Dictionary<string, string> CountryToJurisdiction = new(StringComparer.OrdinalIgnoreCase)
|
||||
{
|
||||
["United Kingdom"] = "gb",
|
||||
["UK"] = "gb",
|
||||
["England"] = "gb",
|
||||
["Scotland"] = "gb",
|
||||
["Wales"] = "gb",
|
||||
["United States"] = "us",
|
||||
["USA"] = "us",
|
||||
["US"] = "us",
|
||||
["Germany"] = "de",
|
||||
["France"] = "fr",
|
||||
["Netherlands"] = "nl",
|
||||
["Ireland"] = "ie",
|
||||
["Spain"] = "es",
|
||||
["Italy"] = "it",
|
||||
["Canada"] = "ca",
|
||||
["Australia"] = "au",
|
||||
["New Zealand"] = "nz",
|
||||
["Singapore"] = "sg",
|
||||
["Hong Kong"] = "hk",
|
||||
["Japan"] = "jp",
|
||||
["India"] = "in",
|
||||
["China"] = "cn",
|
||||
["Brazil"] = "br",
|
||||
["Mexico"] = "mx",
|
||||
["Switzerland"] = "ch",
|
||||
["Sweden"] = "se",
|
||||
["Norway"] = "no",
|
||||
["Denmark"] = "dk",
|
||||
["Finland"] = "fi",
|
||||
["Belgium"] = "be",
|
||||
["Austria"] = "at",
|
||||
["Poland"] = "pl",
|
||||
["Portugal"] = "pt",
|
||||
["UAE"] = "ae",
|
||||
["South Africa"] = "za",
|
||||
};
|
||||
|
||||
public InternationalCompanyVerifierService(
|
||||
OpenCorporatesClient openCorporatesClient,
|
||||
ILogger<InternationalCompanyVerifierService> logger)
|
||||
{
|
||||
_openCorporatesClient = openCorporatesClient;
|
||||
_logger = logger;
|
||||
}
|
||||
|
||||
public async Task<InternationalCompanyResult> VerifyCompanyAsync(
|
||||
string companyName,
|
||||
string? jurisdiction = null,
|
||||
DateOnly? claimedStartDate = null,
|
||||
DateOnly? claimedEndDate = null)
|
||||
{
|
||||
try
|
||||
{
|
||||
_logger.LogInformation("Searching OpenCorporates for: {Company} in {Jurisdiction}",
|
||||
companyName, jurisdiction ?? "all jurisdictions");
|
||||
|
||||
string? jurisdictionCode = null;
|
||||
if (!string.IsNullOrEmpty(jurisdiction) && CountryToJurisdiction.TryGetValue(jurisdiction, out var code))
|
||||
{
|
||||
jurisdictionCode = code;
|
||||
}
|
||||
else if (!string.IsNullOrEmpty(jurisdiction) && jurisdiction.Length == 2)
|
||||
{
|
||||
jurisdictionCode = jurisdiction.ToLowerInvariant();
|
||||
}
|
||||
|
||||
var searchResponse = await _openCorporatesClient.SearchCompaniesAsync(
|
||||
companyName,
|
||||
jurisdictionCode);
|
||||
|
||||
if (searchResponse?.Companies == null || searchResponse.Companies.Count == 0)
|
||||
{
|
||||
return new InternationalCompanyResult
|
||||
{
|
||||
ClaimedCompany = companyName,
|
||||
ClaimedJurisdiction = jurisdiction ?? "Unknown",
|
||||
IsVerified = false,
|
||||
VerificationNotes = "No matching companies found in OpenCorporates"
|
||||
};
|
||||
}
|
||||
|
||||
// Find the best match
|
||||
var bestMatch = FindBestMatch(companyName, searchResponse.Companies);
|
||||
|
||||
if (bestMatch == null)
|
||||
{
|
||||
return new InternationalCompanyResult
|
||||
{
|
||||
ClaimedCompany = companyName,
|
||||
ClaimedJurisdiction = jurisdiction ?? "Unknown",
|
||||
IsVerified = false,
|
||||
VerificationNotes = $"Found {searchResponse.Companies.Count} results but no close name matches"
|
||||
};
|
||||
}
|
||||
|
||||
// Parse dates
|
||||
var incorporationDate = ParseDate(bestMatch.IncorporationDate);
|
||||
var dissolutionDate = ParseDate(bestMatch.DissolutionDate);
|
||||
|
||||
// Calculate match score
|
||||
var matchScore = CalculateMatchScore(companyName, bestMatch.Name ?? "");
|
||||
|
||||
// Check for timeline issues
|
||||
var flags = new List<CompanyVerificationFlag>();
|
||||
|
||||
if (claimedStartDate.HasValue && incorporationDate.HasValue &&
|
||||
claimedStartDate.Value < incorporationDate.Value)
|
||||
{
|
||||
flags.Add(new CompanyVerificationFlag
|
||||
{
|
||||
Type = "EmploymentBeforeIncorporation",
|
||||
Severity = "Critical",
|
||||
Message = $"Claimed start date ({claimedStartDate:yyyy-MM-dd}) is before company incorporation ({incorporationDate:yyyy-MM-dd})",
|
||||
ScoreImpact = -30
|
||||
});
|
||||
}
|
||||
|
||||
if (claimedEndDate.HasValue && dissolutionDate.HasValue &&
|
||||
claimedEndDate.Value > dissolutionDate.Value)
|
||||
{
|
||||
flags.Add(new CompanyVerificationFlag
|
||||
{
|
||||
Type = "EmploymentAfterDissolution",
|
||||
Severity = "Warning",
|
||||
Message = $"Claimed end date ({claimedEndDate:yyyy-MM-dd}) is after company dissolution ({dissolutionDate:yyyy-MM-dd})",
|
||||
ScoreImpact = -20
|
||||
});
|
||||
}
|
||||
|
||||
return new InternationalCompanyResult
|
||||
{
|
||||
ClaimedCompany = companyName,
|
||||
ClaimedJurisdiction = jurisdiction ?? "Unknown",
|
||||
IsVerified = true,
|
||||
MatchedCompanyName = bestMatch.Name,
|
||||
CompanyNumber = bestMatch.CompanyNumber,
|
||||
Jurisdiction = GetJurisdictionName(bestMatch.JurisdictionCode),
|
||||
JurisdictionCode = bestMatch.JurisdictionCode,
|
||||
CompanyType = bestMatch.CompanyType,
|
||||
Status = bestMatch.CurrentStatus,
|
||||
IncorporationDate = incorporationDate,
|
||||
DissolutionDate = dissolutionDate,
|
||||
RegisteredAddress = bestMatch.RegisteredAddressInFull,
|
||||
OpenCorporatesUrl = bestMatch.OpencorporatesUrl,
|
||||
DataLastUpdated = bestMatch.UpdatedAt,
|
||||
MatchScore = matchScore,
|
||||
VerificationNotes = BuildVerificationSummary(bestMatch, searchResponse.TotalCount),
|
||||
Flags = flags
|
||||
};
|
||||
}
|
||||
catch (Exception ex)
|
||||
{
|
||||
_logger.LogError(ex, "Error searching OpenCorporates for: {Company}", companyName);
|
||||
return new InternationalCompanyResult
|
||||
{
|
||||
ClaimedCompany = companyName,
|
||||
ClaimedJurisdiction = jurisdiction ?? "Unknown",
|
||||
IsVerified = false,
|
||||
VerificationNotes = $"Error during search: {ex.Message}"
|
||||
};
|
||||
}
|
||||
}
|
||||
|
||||
public async Task<List<OpenCorporatesSearchResult>> SearchCompaniesAsync(
|
||||
string query,
|
||||
string? jurisdiction = null)
|
||||
{
|
||||
try
|
||||
{
|
||||
string? jurisdictionCode = null;
|
||||
if (!string.IsNullOrEmpty(jurisdiction) && CountryToJurisdiction.TryGetValue(jurisdiction, out var code))
|
||||
{
|
||||
jurisdictionCode = code;
|
||||
}
|
||||
|
||||
var searchResponse = await _openCorporatesClient.SearchCompaniesAsync(
|
||||
query,
|
||||
jurisdictionCode);
|
||||
|
||||
if (searchResponse?.Companies == null)
|
||||
{
|
||||
return [];
|
||||
}
|
||||
|
||||
return searchResponse.Companies
|
||||
.Where(c => c.Company != null)
|
||||
.Select(c => new OpenCorporatesSearchResult
|
||||
{
|
||||
CompanyName = c.Company!.Name ?? "Unknown",
|
||||
CompanyNumber = c.Company.CompanyNumber ?? "Unknown",
|
||||
Jurisdiction = GetJurisdictionName(c.Company.JurisdictionCode),
|
||||
JurisdictionCode = c.Company.JurisdictionCode,
|
||||
Status = c.Company.CurrentStatus,
|
||||
IncorporationDate = ParseDate(c.Company.IncorporationDate),
|
||||
OpenCorporatesUrl = c.Company.OpencorporatesUrl,
|
||||
MatchScore = CalculateMatchScore(query, c.Company.Name ?? "")
|
||||
})
|
||||
.ToList();
|
||||
}
|
||||
catch (Exception ex)
|
||||
{
|
||||
_logger.LogError(ex, "Error searching OpenCorporates for: {Query}", query);
|
||||
return [];
|
||||
}
|
||||
}
|
||||
|
||||
public async Task<List<JurisdictionInfo>> GetJurisdictionsAsync()
|
||||
{
|
||||
try
|
||||
{
|
||||
var jurisdictions = await _openCorporatesClient.GetJurisdictionsAsync();
|
||||
|
||||
if (jurisdictions == null)
|
||||
{
|
||||
return [];
|
||||
}
|
||||
|
||||
return jurisdictions
|
||||
.Select(j => new JurisdictionInfo
|
||||
{
|
||||
Code = j.Code ?? "Unknown",
|
||||
Name = j.FullName ?? j.Name ?? "Unknown",
|
||||
Country = j.Country
|
||||
})
|
||||
.ToList();
|
||||
}
|
||||
catch (Exception ex)
|
||||
{
|
||||
_logger.LogError(ex, "Error getting OpenCorporates jurisdictions");
|
||||
return [];
|
||||
}
|
||||
}
|
||||
|
||||
private static OpenCorporatesCompany? FindBestMatch(
|
||||
string searchName,
|
||||
List<OpenCorporatesCompanyWrapper> companies)
|
||||
{
|
||||
var normalizedSearch = NormalizeName(searchName);
|
||||
|
||||
// First try exact match
|
||||
var exactMatch = companies
|
||||
.Select(c => c.Company)
|
||||
.Where(c => c != null)
|
||||
.FirstOrDefault(c => NormalizeName(c!.Name ?? "").Equals(normalizedSearch, StringComparison.OrdinalIgnoreCase));
|
||||
|
||||
if (exactMatch != null)
|
||||
return exactMatch;
|
||||
|
||||
// Then try contains match, preferring active companies
|
||||
var containsMatches = companies
|
||||
.Select(c => c.Company)
|
||||
.Where(c => c != null && !string.IsNullOrEmpty(c.Name))
|
||||
.Where(c => NormalizeName(c!.Name!).Contains(normalizedSearch, StringComparison.OrdinalIgnoreCase) ||
|
||||
normalizedSearch.Contains(NormalizeName(c!.Name!), StringComparison.OrdinalIgnoreCase))
|
||||
.OrderBy(c => c!.CurrentStatus?.ToLowerInvariant() == "active" ? 0 : 1)
|
||||
.ThenBy(c => c!.Name!.Length)
|
||||
.ToList();
|
||||
|
||||
return containsMatches.FirstOrDefault();
|
||||
}
|
||||
|
||||
private static string NormalizeName(string name)
|
||||
{
|
||||
return name
|
||||
.Replace("LIMITED", "", StringComparison.OrdinalIgnoreCase)
|
||||
.Replace("LTD", "", StringComparison.OrdinalIgnoreCase)
|
||||
.Replace("PLC", "", StringComparison.OrdinalIgnoreCase)
|
||||
.Replace("INC", "", StringComparison.OrdinalIgnoreCase)
|
||||
.Replace("CORP", "", StringComparison.OrdinalIgnoreCase)
|
||||
.Replace("LLC", "", StringComparison.OrdinalIgnoreCase)
|
||||
.Replace("GMBH", "", StringComparison.OrdinalIgnoreCase)
|
||||
.Replace("AG", "", StringComparison.OrdinalIgnoreCase)
|
||||
.Replace(".", "")
|
||||
.Replace(",", "")
|
||||
.Trim();
|
||||
}
|
||||
|
||||
private static string GetJurisdictionName(string? code)
|
||||
{
|
||||
if (string.IsNullOrEmpty(code))
|
||||
return "Unknown";
|
||||
|
||||
return code.ToUpperInvariant();
|
||||
}
|
||||
|
||||
private static DateOnly? ParseDate(string? dateString)
|
||||
{
|
||||
if (string.IsNullOrEmpty(dateString))
|
||||
return null;
|
||||
|
||||
if (DateOnly.TryParse(dateString, out var date))
|
||||
return date;
|
||||
|
||||
return null;
|
||||
}
|
||||
|
||||
private static int CalculateMatchScore(string searchName, string foundName)
|
||||
{
|
||||
var normalizedSearch = NormalizeName(searchName).ToLowerInvariant();
|
||||
var normalizedFound = NormalizeName(foundName).ToLowerInvariant();
|
||||
|
||||
if (normalizedSearch == normalizedFound)
|
||||
return 100;
|
||||
|
||||
if (normalizedFound.Contains(normalizedSearch) || normalizedSearch.Contains(normalizedFound))
|
||||
return 80;
|
||||
|
||||
// Calculate Levenshtein similarity
|
||||
var distance = LevenshteinDistance(normalizedSearch, normalizedFound);
|
||||
var maxLength = Math.Max(normalizedSearch.Length, normalizedFound.Length);
|
||||
var similarity = (int)((1 - (double)distance / maxLength) * 100);
|
||||
|
||||
return Math.Max(0, similarity);
|
||||
}
|
||||
|
||||
private static int LevenshteinDistance(string s1, string s2)
|
||||
{
|
||||
var n = s1.Length;
|
||||
var m = s2.Length;
|
||||
var d = new int[n + 1, m + 1];
|
||||
|
||||
for (var i = 0; i <= n; i++)
|
||||
d[i, 0] = i;
|
||||
|
||||
for (var j = 0; j <= m; j++)
|
||||
d[0, j] = j;
|
||||
|
||||
for (var i = 1; i <= n; i++)
|
||||
{
|
||||
for (var j = 1; j <= m; j++)
|
||||
{
|
||||
var cost = s1[i - 1] == s2[j - 1] ? 0 : 1;
|
||||
d[i, j] = Math.Min(Math.Min(d[i - 1, j] + 1, d[i, j - 1] + 1), d[i - 1, j - 1] + cost);
|
||||
}
|
||||
}
|
||||
|
||||
return d[n, m];
|
||||
}
|
||||
|
||||
private static string BuildVerificationSummary(OpenCorporatesCompany company, int totalResults)
|
||||
{
|
||||
var parts = new List<string>();
|
||||
|
||||
if (!string.IsNullOrEmpty(company.CurrentStatus))
|
||||
parts.Add($"Status: {company.CurrentStatus}");
|
||||
|
||||
if (!string.IsNullOrEmpty(company.JurisdictionCode))
|
||||
parts.Add($"Jurisdiction: {company.JurisdictionCode.ToUpperInvariant()}");
|
||||
|
||||
if (!string.IsNullOrEmpty(company.IncorporationDate))
|
||||
parts.Add($"Incorporated: {company.IncorporationDate}");
|
||||
|
||||
if (!string.IsNullOrEmpty(company.CompanyType))
|
||||
parts.Add($"Type: {company.CompanyType}");
|
||||
|
||||
if (totalResults > 1)
|
||||
parts.Add($"({totalResults} total matches found)");
|
||||
|
||||
return string.Join(" | ", parts);
|
||||
}
|
||||
}
|
||||
Reference in New Issue
Block a user