Move brewery reverse geocoding from C# seeding to C++ pipeline (#249)

This commit is contained in:
2026-08-25 21:55:34 -04:00
committed by Aaron Po
parent 2b82d1918b
commit ad1c09f139
25 changed files with 377 additions and 185 deletions

View File

@@ -144,6 +144,8 @@ FetchContent_MakeAvailable(cpp-httplib)
# 5. Executable & Sources # 5. Executable & Sources
add_executable(${PROJECT_NAME} add_executable(${PROJECT_NAME}
includes/services/enrichment/mock_enrichment.h includes/services/enrichment/mock_enrichment.h
includes/services/address/address_service.h
includes/services/address/mock_address_service.h
includes/services/curated_data/mock_curated_data_service.h includes/services/curated_data/mock_curated_data_service.h
includes/json_handling/pretty_print.h) includes/json_handling/pretty_print.h)
@@ -225,6 +227,11 @@ target_sources(${PROJECT_NAME} PRIVATE
src/services/enrichment/wikipedia/get_summary.cc src/services/enrichment/wikipedia/get_summary.cc
) )
# --- services: address ---
target_sources(${PROJECT_NAME} PRIVATE
src/services/address/nominatim_address_service.cc
)
# --- services: xeger --- # --- services: xeger ---
target_sources(${PROJECT_NAME} PRIVATE target_sources(${PROJECT_NAME} PRIVATE
src/services/xeger/xeger.cc src/services/xeger/xeger.cc

View File

@@ -13,6 +13,11 @@
// --- concurrency --- // --- concurrency ---
#include "concurrency/bounded_channel.h" #include "concurrency/bounded_channel.h"
// --- services: address ---
#include "services/address/address_service.h"
#include "services/address/mock_address_service.h"
#include "services/address/nominatim_address_service.h"
// --- data_generation --- // --- data_generation ---
#include "data_generation/data_generator.h" #include "data_generation/data_generator.h"
#include "data_generation/generation_json_validation.h" #include "data_generation/generation_json_validation.h"

View File

@@ -12,6 +12,7 @@
#include "data_generation/data_generator.h" #include "data_generation/data_generator.h"
#include "data_model/generated_models.h" #include "data_model/generated_models.h"
#include "services/address/address_service.h"
#include "services/curated_data/curated_data_service.h" #include "services/curated_data/curated_data_service.h"
#include "services/database/export_service.h" #include "services/database/export_service.h"
#include "services/enrichment/enrichment_service.h" #include "services/enrichment/enrichment_service.h"
@@ -34,6 +35,8 @@ class BiergartenPipelineOrchestrator {
* @param exporter Database backend for persisting generated records. * @param exporter Database backend for persisting generated records.
* @param curated_data_service Loads curated location, persona, and name * @param curated_data_service Loads curated location, persona, and name
* data used to seed generation. * data used to seed generation.
* @param address_service Resolves a street address for generated
* brewery coordinates.
* @param application_options CLI configuration and paths. * @param application_options CLI configuration and paths.
*/ */
BiergartenPipelineOrchestrator( BiergartenPipelineOrchestrator(
@@ -42,6 +45,7 @@ class BiergartenPipelineOrchestrator {
std::unique_ptr<DataGenerator> generator, std::unique_ptr<DataGenerator> generator,
std::unique_ptr<IExportService> exporter, std::unique_ptr<IExportService> exporter,
std::unique_ptr<ICuratedDataService> curated_data_service, std::unique_ptr<ICuratedDataService> curated_data_service,
std::unique_ptr<IAddressService> address_service,
const ApplicationOptions& application_options); const ApplicationOptions& application_options);
/** /**
@@ -78,6 +82,7 @@ class BiergartenPipelineOrchestrator {
*/ */
std::unique_ptr<IExportService> exporter_; std::unique_ptr<IExportService> exporter_;
std::unique_ptr<ICuratedDataService> curated_data_service_; std::unique_ptr<ICuratedDataService> curated_data_service_;
std::unique_ptr<IAddressService> address_service_;
ApplicationOptions application_options_; ApplicationOptions application_options_;

View File

@@ -95,6 +95,18 @@ struct BreweryAddress {
City city; City city;
double longitude{}; double longitude{};
double latitude{}; double latitude{};
/**
* @brief Street address line (e.g. house number + road), resolved once
* via reverse geocoding at generation time.
*/
std::string address_line1{};
/**
* @brief Postal code, resolved once via reverse geocoding at generation
* time.
*/
std::string postal_code{};
}; };
/** /**

View File

@@ -0,0 +1,40 @@
#ifndef BIERGARTEN_PIPELINE_INCLUDES_SERVICES_ADDRESS_ADDRESS_SERVICE_H_
#define BIERGARTEN_PIPELINE_INCLUDES_SERVICES_ADDRESS_ADDRESS_SERVICE_H_
/**
* @file services/address/address_service.h
* @brief Abstraction for resolving a street address from coordinates.
*/
#include <optional>
#include <string>
/**
* @brief Street-level address resolved from a coordinate pair.
*/
struct AddressLookupResult {
std::string address_line1;
std::string postal_code;
};
/**
* @brief Interface for services that reverse-geocode coordinates into a
* street address.
*/
class IAddressService {
public:
virtual ~IAddressService() = default;
/**
* @brief Reverse-geocodes a coordinate pair into a street address.
*
* @param longitude Longitude of the point to resolve.
* @param latitude Latitude of the point to resolve.
* @return The resolved address, or std::nullopt if it could not be
* resolved.
*/
virtual std::optional<AddressLookupResult> ReverseGeocode(
double longitude, double latitude) = 0;
};
#endif // BIERGARTEN_PIPELINE_INCLUDES_SERVICES_ADDRESS_ADDRESS_SERVICE_H_

View File

@@ -0,0 +1,27 @@
#ifndef BIERGARTEN_PIPELINE_INCLUDES_SERVICES_ADDRESS_MOCK_ADDRESS_SERVICE_H_
#define BIERGARTEN_PIPELINE_INCLUDES_SERVICES_ADDRESS_MOCK_ADDRESS_SERVICE_H_
/**
* @file services/address/mock_address_service.h
* @brief Deterministic IAddressService used when network geocoding is
* disabled.
*/
#include <optional>
#include "services/address/address_service.h"
/**
* @brief Address service that returns a fixed placeholder address for any
* coordinate pair, without making network calls.
*/
class MockAddressService final : public IAddressService {
public:
std::optional<AddressLookupResult> ReverseGeocode(
double /*longitude*/, double /*latitude*/) override {
return AddressLookupResult{.address_line1 = "123 Mock Street",
.postal_code = "00000"};
}
};
#endif // BIERGARTEN_PIPELINE_INCLUDES_SERVICES_ADDRESS_MOCK_ADDRESS_SERVICE_H_

View File

@@ -0,0 +1,37 @@
#ifndef BIERGARTEN_PIPELINE_INCLUDES_SERVICES_ADDRESS_NOMINATIM_ADDRESS_SERVICE_H_
#define BIERGARTEN_PIPELINE_INCLUDES_SERVICES_ADDRESS_NOMINATIM_ADDRESS_SERVICE_H_
/**
* @file services/address/nominatim_address_service.h
* @brief Reverse geocoding backed by the public Nominatim API.
*/
#include <memory>
#include <optional>
#include "services/address/address_service.h"
#include "services/logging/logger.h"
#include "web_client/web_client.h"
/**
* @brief Reverse-geocodes coordinates into street addresses via the public
* Nominatim API (https://nominatim.openstreetmap.org).
*
* Required for API usage:
* - single identifying User-Agent and at most one request per
* second (https://operations.osmfoundation.org/policies/nominatim/).
*/
class NominatimAddressService final : public IAddressService {
public:
explicit NominatimAddressService(std::unique_ptr<WebClient> client,
std::shared_ptr<ILogger> logger);
std::optional<AddressLookupResult> ReverseGeocode(
double longitude, double latitude) override;
private:
std::unique_ptr<WebClient> client_;
std::shared_ptr<ILogger> logger_;
};
#endif // BIERGARTEN_PIPELINE_INCLUDES_SERVICES_ADDRESS_NOMINATIM_ADDRESS_SERVICE_H_

View File

@@ -56,6 +56,8 @@ CREATE TABLE IF NOT EXISTS brewery_addresses (
city_id INTEGER NOT NULL, city_id INTEGER NOT NULL,
longitude REAL NOT NULL, longitude REAL NOT NULL,
latitude REAL NOT NULL, latitude REAL NOT NULL,
address_line1 TEXT NOT NULL,
postal_code TEXT NOT NULL,
FOREIGN KEY(brewery_id) REFERENCES breweries(id) ON DELETE CASCADE, FOREIGN KEY(brewery_id) REFERENCES breweries(id) ON DELETE CASCADE,
FOREIGN KEY(city_id) REFERENCES cities(id) ON DELETE CASCADE FOREIGN KEY(city_id) REFERENCES cities(id) ON DELETE CASCADE
); );
@@ -131,8 +133,10 @@ INSERT INTO brewery_addresses (
brewery_id, brewery_id,
city_id, city_id,
longitude, longitude,
latitude latitude,
) VALUES (?, ?, ?, ?); address_line1,
postal_code
) VALUES (?, ?, ?, ?, ?, ?);
)sql"; )sql";
inline constexpr std::string_view kInsertUserSql = R"sql( inline constexpr std::string_view kInsertUserSql = R"sql(
@@ -184,6 +188,8 @@ enum BreweryAddressBindIndex {
kBreweryAddressCityIdBindIndex, kBreweryAddressCityIdBindIndex,
kBreweryAddressLongitudeBindIndex, kBreweryAddressLongitudeBindIndex,
kBreweryAddressLatitudeBindIndex, kBreweryAddressLatitudeBindIndex,
kBreweryAddressLine1BindIndex,
kBreweryAddressPostalCodeBindIndex,
}; };
enum UserBindIndex { enum UserBindIndex {

View File

@@ -13,10 +13,12 @@ BiergartenPipelineOrchestrator::BiergartenPipelineOrchestrator(
std::unique_ptr<DataGenerator> generator, std::unique_ptr<DataGenerator> generator,
std::unique_ptr<IExportService> exporter, std::unique_ptr<IExportService> exporter,
std::unique_ptr<ICuratedDataService> curated_data_service, std::unique_ptr<ICuratedDataService> curated_data_service,
std::unique_ptr<IAddressService> address_service,
const ApplicationOptions& application_options) const ApplicationOptions& application_options)
: logger_(std::move(logger)), : logger_(std::move(logger)),
context_service_(std::move(context_service)), context_service_(std::move(context_service)),
generator_(std::move(generator)), generator_(std::move(generator)),
exporter_(std::move(exporter)), exporter_(std::move(exporter)),
curated_data_service_(std::move(curated_data_service)), curated_data_service_(std::move(curated_data_service)),
address_service_(std::move(address_service)),
application_options_(application_options) {} application_options_(application_options) {}

View File

@@ -7,14 +7,21 @@
#include <format> #include <format>
#include <optional> #include <optional>
#include <random> #include <random>
#include <string>
#include <string_view>
#include "biergarten_pipeline_orchestrator.h" #include "biergarten_pipeline_orchestrator.h"
#include "biergarten_pipeline_orchestrator/coords.h" #include "biergarten_pipeline_orchestrator/coords.h"
#include "services/address/address_service.h"
#include "services/logging/logger.h" #include "services/logging/logger.h"
using biergarten_pipeline_orchestrator_internal::Coords; using biergarten_pipeline_orchestrator_internal::Coords;
using biergarten_pipeline_orchestrator_internal::GetRandomCoordsWithinRange; using biergarten_pipeline_orchestrator_internal::GetRandomCoordsWithinRange;
namespace {
constexpr std::string_view kPlaceholderAddressLine1 = "Address unavailable";
} // namespace
void BiergartenPipelineOrchestrator::GenerateBreweries( void BiergartenPipelineOrchestrator::GenerateBreweries(
std::span<const EnrichedCity> cities) { std::span<const EnrichedCity> cities) {
logger_->Log({.level = LogLevel::Info, logger_->Log({.level = LogLevel::Info,
@@ -39,10 +46,32 @@ void BiergartenPipelineOrchestrator::GenerateBreweries(
const Coords brewery_coords = GetRandomCoordsWithinRange( const Coords brewery_coords = GetRandomCoordsWithinRange(
city_centre, kMaxDistanceFromCentreKm, rng); city_centre, kMaxDistanceFromCentreKm, rng);
const std::optional<AddressLookupResult> geocoded =
address_service_->ReverseGeocode(brewery_coords.longitude,
brewery_coords.latitude);
if (!geocoded.has_value()) {
logger_->Log(
{.level = LogLevel::Warn,
.phase = PipelinePhase::BreweryAndBeerGeneration,
.message = std::format(
"[Pipeline] Could not reverse geocode brewery in '{}' "
"({}, {}); using a placeholder address.",
enriched_city.location.city, brewery_coords.latitude,
brewery_coords.longitude)});
}
return BreweryRecord{ return BreweryRecord{
.address = BreweryAddress{.city = enriched_city.location, .address =
BreweryAddress{
.city = enriched_city.location,
.longitude = brewery_coords.longitude, .longitude = brewery_coords.longitude,
.latitude = brewery_coords.latitude}, .latitude = brewery_coords.latitude,
.address_line1 = geocoded.has_value()
? geocoded->address_line1
: std::string(kPlaceholderAddressLine1),
.postal_code =
geocoded.has_value() ? geocoded->postal_code : ""},
.brewery = brewery}; .brewery = brewery};
} catch (const std::exception& e) { } catch (const std::exception& e) {
++skipped_count; ++skipped_count;

View File

@@ -166,6 +166,26 @@ int main(const int argc, char** argv) {
inj.template create<std::unique_ptr<WebClient>>(), inj.template create<std::unique_ptr<WebClient>>(),
log_producer); log_producer);
}), }),
di::bind<IAddressService>().to(
[options, &log_producer](
const auto& inj) -> std::unique_ptr<IAddressService> {
if (options.generator.mode == GeneratorMode::kMock) {
log_producer->Log({.level = LogLevel::Info,
.phase = PipelinePhase::Startup,
.message = "Address service: mock"});
return std::make_unique<MockAddressService>();
}
log_producer->Log(
{.level = LogLevel::Info,
.phase = PipelinePhase::Startup,
.message = "Address service: Nominatim"});
return std::make_unique<NominatimAddressService>(
inj.template create<std::unique_ptr<WebClient>>(),
log_producer);
}),
di::bind<DataGenerator>().to([&options, &model_path, &sampling, di::bind<DataGenerator>().to([&options, &model_path, &sampling,
&prompt_directory, &prompt_directory,
&log_producer](const auto& inj) &log_producer](const auto& inj)

View File

@@ -0,0 +1,133 @@
/**
* @file services/address/nominatim_address_service.cc
* @brief NominatimAddressService implementation.
*/
#include "services/address/nominatim_address_service.h"
#include <boost/json.hpp>
#include <chrono>
#include <format>
#include <stdexcept>
#include <string>
#include <thread>
#include <utility>
using namespace boost;
NominatimAddressService::NominatimAddressService(
std::unique_ptr<WebClient> client, std::shared_ptr<ILogger> logger)
: client_(std::move(client)), logger_(std::move(logger)) {}
std::optional<AddressLookupResult> NominatimAddressService::ReverseGeocode(
double longitude, double latitude) {
if (!client_) {
if (logger_) {
logger_->Log({.level = LogLevel::Warn,
.phase = PipelinePhase::Enrichment,
.message = "Nominatim client is nullptr."});
}
return std::nullopt;
}
const std::string url = std::format(
"https://nominatim.openstreetmap.org/reverse?format=jsonv2&lat={}&"
"lon={}&zoom=18&addressdetails=1",
latitude, longitude);
std::string body;
try {
body = client_->Get(url);
} catch (const std::runtime_error& e) {
if (logger_) {
logger_->Log({.level = LogLevel::Warn,
.phase = PipelinePhase::Enrichment,
.message = std::format(
"NominatimAddressService: request failed for "
"({}, {}): {}",
latitude, longitude, e.what())});
}
return std::nullopt;
}
{
using namespace std::literals::chrono_literals;
std::this_thread::sleep_for(1s);
}
system::error_code ec;
const json::value doc = json::parse(body, ec);
if (ec) {
if (logger_) {
logger_->Log({.level = LogLevel::Warn,
.phase = PipelinePhase::Enrichment,
.message = std::format(
"NominatimAddressService: JSON parse error for "
"({}, {}): {}",
latitude, longitude, ec.message())});
}
return std::nullopt;
}
const json::object* obj = doc.if_object();
if ((obj == nullptr) || obj->contains("error")) {
if (logger_) {
logger_->Log(
{.level = LogLevel::Warn,
.phase = PipelinePhase::Enrichment,
.message = std::format("NominatimAddressService: no address "
"found for ({}, {})",
latitude, longitude)});
}
return std::nullopt;
}
const json::value* address_ptr = obj->if_contains("address");
if ((address_ptr == nullptr) || !address_ptr->is_object()) {
if (logger_) {
logger_->Log(
{.level = LogLevel::Warn,
.phase = PipelinePhase::Enrichment,
.message = std::format("NominatimAddressService: missing "
"'address' for ({}, {})",
latitude, longitude)});
}
return std::nullopt;
}
const json::object& address = address_ptr->get_object();
const auto extract_string = [&address](std::string_view key) -> std::string {
const json::value* value_ptr = address.if_contains(key);
if ((value_ptr == nullptr) || !value_ptr->is_string()) {
return {};
}
return std::string(value_ptr->get_string());
};
const std::string house_number = extract_string("house_number");
const std::string road = extract_string("road");
const std::string postal_code = extract_string("postcode");
std::string address_line1 = road;
if (!house_number.empty()) {
address_line1 = address_line1.empty() ? house_number
: house_number + " " + road;
}
if (address_line1.empty()) {
if (logger_) {
logger_->Log(
{.level = LogLevel::Warn,
.phase = PipelinePhase::Enrichment,
.message = std::format("NominatimAddressService: no usable "
"street address for ({}, {})",
latitude, longitude)});
}
return std::nullopt;
}
return AddressLookupResult{.address_line1 = address_line1,
.postal_code = postal_code};
}

View File

@@ -184,6 +184,20 @@ uint64_t SqliteExportService::ProcessRecord(const BreweryRecord& brewery) {
sqlite_export_service_internal::kBreweryAddressLatitudeBindIndex, sqlite_export_service_internal::kBreweryAddressLatitudeBindIndex,
.value = brewery.address.latitude, .value = brewery.address.latitude,
.action = "Failed to bind SQLite brewery address latitude"}); .action = "Failed to bind SQLite brewery address latitude"});
sqlite_export_service_internal::Bind(
insert_brewery_address_stmt_,
sqlite_export_service_internal::BoundParam<std::string_view>{
.index =
sqlite_export_service_internal::kBreweryAddressLine1BindIndex,
.value = brewery.address.address_line1,
.action = "Failed to bind SQLite brewery address line 1"});
sqlite_export_service_internal::Bind(
insert_brewery_address_stmt_,
sqlite_export_service_internal::BoundParam<std::string_view>{
.index = sqlite_export_service_internal::
kBreweryAddressPostalCodeBindIndex,
.value = brewery.address.postal_code,
.action = "Failed to bind SQLite brewery address postal code"});
sqlite_export_service_internal::StepStatement( sqlite_export_service_internal::StepStatement(
db_handle_, insert_brewery_address_stmt_, db_handle_, insert_brewery_address_stmt_,

View File

@@ -32,6 +32,7 @@ COPY ["Features/Features.Locations/Features.Locations.csproj", "Features/Feature
COPY ["Features/Features.Locations.Tests/Features.Locations.Tests.csproj", "Features/Features.Locations.Tests/"] COPY ["Features/Features.Locations.Tests/Features.Locations.Tests.csproj", "Features/Features.Locations.Tests/"]
COPY ["Features/Features.UserManagement/Features.UserManagement.csproj", "Features/Features.UserManagement/"] COPY ["Features/Features.UserManagement/Features.UserManagement.csproj", "Features/Features.UserManagement/"]
COPY ["Features/Features.UserManagement.Tests/Features.UserManagement.Tests.csproj", "Features/Features.UserManagement.Tests/"] COPY ["Features/Features.UserManagement.Tests/Features.UserManagement.Tests.csproj", "Features/Features.UserManagement.Tests/"]
COPY ["Infrastructure/Infrastructure.Configuration/Infrastructure.Configuration.csproj", "Infrastructure/Infrastructure.Configuration/"]
COPY ["Infrastructure/Infrastructure.Email/Infrastructure.Email.csproj", "Infrastructure/Infrastructure.Email/"] COPY ["Infrastructure/Infrastructure.Email/Infrastructure.Email.csproj", "Infrastructure/Infrastructure.Email/"]
COPY ["Infrastructure/Infrastructure.Email.Templates/Infrastructure.Email.Templates.csproj", "Infrastructure/Infrastructure.Email.Templates/"] COPY ["Infrastructure/Infrastructure.Email.Templates/Infrastructure.Email.Templates.csproj", "Infrastructure/Infrastructure.Email.Templates/"]
COPY ["Infrastructure/Infrastructure.Jwt/Infrastructure.Jwt.csproj", "Infrastructure/Infrastructure.Jwt/"] COPY ["Infrastructure/Infrastructure.Jwt/Infrastructure.Jwt.csproj", "Infrastructure/Infrastructure.Jwt/"]

View File

@@ -22,6 +22,7 @@ COPY ["Features/Features.Locations/Features.Locations.csproj", "Features/Feature
COPY ["Features/Features.Locations.Tests/Features.Locations.Tests.csproj", "Features/Features.Locations.Tests/"] COPY ["Features/Features.Locations.Tests/Features.Locations.Tests.csproj", "Features/Features.Locations.Tests/"]
COPY ["Features/Features.UserManagement/Features.UserManagement.csproj", "Features/Features.UserManagement/"] COPY ["Features/Features.UserManagement/Features.UserManagement.csproj", "Features/Features.UserManagement/"]
COPY ["Features/Features.UserManagement.Tests/Features.UserManagement.Tests.csproj", "Features/Features.UserManagement.Tests/"] COPY ["Features/Features.UserManagement.Tests/Features.UserManagement.Tests.csproj", "Features/Features.UserManagement.Tests/"]
COPY ["Infrastructure/Infrastructure.Configuration/Infrastructure.Configuration.csproj", "Infrastructure/Infrastructure.Configuration/"]
COPY ["Infrastructure/Infrastructure.Email/Infrastructure.Email.csproj", "Infrastructure/Infrastructure.Email/"] COPY ["Infrastructure/Infrastructure.Email/Infrastructure.Email.csproj", "Infrastructure/Infrastructure.Email/"]
COPY ["Infrastructure/Infrastructure.Email.Templates/Infrastructure.Email.Templates.csproj", "Infrastructure/Infrastructure.Email.Templates/"] COPY ["Infrastructure/Infrastructure.Email.Templates/Infrastructure.Email.Templates.csproj", "Infrastructure/Infrastructure.Email.Templates/"]
COPY ["Infrastructure/Infrastructure.Jwt/Infrastructure.Jwt.csproj", "Infrastructure/Infrastructure.Jwt/"] COPY ["Infrastructure/Infrastructure.Jwt/Infrastructure.Jwt.csproj", "Infrastructure/Infrastructure.Jwt/"]

View File

@@ -13,6 +13,7 @@
<Project Path="Domain/Domain.Exceptions/Domain.Exceptions.csproj" /> <Project Path="Domain/Domain.Exceptions/Domain.Exceptions.csproj" />
</Folder> </Folder>
<Folder Name="/Infrastructure/"> <Folder Name="/Infrastructure/">
<Project Path="Infrastructure/Infrastructure.Configuration/Infrastructure.Configuration.csproj" />
<Project Path="Infrastructure/Infrastructure.Email/Infrastructure.Email.csproj" /> <Project Path="Infrastructure/Infrastructure.Email/Infrastructure.Email.csproj" />
<Project Path="Infrastructure/Infrastructure.Email.Templates/Infrastructure.Email.Templates.csproj" /> <Project Path="Infrastructure/Infrastructure.Email.Templates/Infrastructure.Email.Templates.csproj" />
<Project Path="Infrastructure/Infrastructure.Jwt/Infrastructure.Jwt.csproj" /> <Project Path="Infrastructure/Infrastructure.Jwt/Infrastructure.Jwt.csproj" />

View File

@@ -24,7 +24,7 @@
<ItemGroup> <ItemGroup>
<None <None
CopyToOutputDirectory="PreserveNewest" CopyToOutputDirectory="PreserveNewest"
Include="SeedData\biergarten_seed_2026-08-24T01-52-09.841693Z.sqlite" Include="SeedData\biergarten_seed_2026-08-25T20-45-50.697244Z.sqlite"
/> />
</ItemGroup> </ItemGroup>
</Project> </Project>

View File

@@ -20,5 +20,5 @@ public static class ConnectionStrings
); );
public static string SqliteConnectionString => public static string SqliteConnectionString =>
"Data Source=SeedData/biergarten_seed_2026-08-24T01-52-09.841693Z.sqlite"; "Data Source=SeedData/biergarten_seed_2026-08-25T20-45-50.697244Z.sqlite";
} }

View File

@@ -1,124 +0,0 @@
using System.Globalization;
using System.Net.Http.Json;
using System.Text.Json.Serialization;
namespace Database.Seed.Geocoding;
public sealed record ReverseGeocodeResult(string AddressLine1, string PostalCode);
internal sealed record NominatimAddress
{
[JsonPropertyName("house_number")]
public string? HouseNumber { get; init; }
[JsonPropertyName("road")]
public string? Road { get; init; }
[JsonPropertyName("postcode")]
public string? Postcode { get; init; }
}
internal sealed record NominatimReverseGeocodeResponse
{
[JsonPropertyName("address")]
public NominatimAddress? Address { get; init; }
[JsonPropertyName("error")]
public string? Error { get; init; }
}
/// <summary>
/// Reverse geocodes coordinates into street addresses via the public Nominatim API
/// (https://nominatim.openstreetmap.org), following its usage policy of a single
/// identifying User-Agent and at most one request per second
/// (https://operations.osmfoundation.org/policies/nominatim/).
/// </summary>
public sealed class NominatimReverseGeocoder : IDisposable
{
private const string ReverseEndpoint = "https://nominatim.openstreetmap.org/reverse";
private static readonly TimeSpan MinRequestInterval = TimeSpan.FromSeconds(1);
private readonly HttpClient _httpClient;
private readonly SemaphoreSlim _rateLimitLock = new(1, 1);
private DateTime _lastRequestUtc = DateTime.MinValue;
public NominatimReverseGeocoder(string userAgent)
{
_httpClient = new HttpClient();
_httpClient.DefaultRequestHeaders.UserAgent.ParseAdd(userAgent);
}
/// <summary>
/// Reverse geocodes a coordinate pair, returning <c>null</c> if Nominatim has no address
/// for the point or the request fails so callers can fall back to a placeholder address.
/// </summary>
public async Task<ReverseGeocodeResult?> ReverseGeocodeAsync(
double longitude,
double latitude,
CancellationToken cancellationToken = default
)
{
await WaitForRateLimitAsync(cancellationToken);
string url =
$"{ReverseEndpoint}?format=jsonv2"
+ $"&lat={latitude.ToString(CultureInfo.InvariantCulture)}"
+ $"&lon={longitude.ToString(CultureInfo.InvariantCulture)}"
+ "&zoom=18&addressdetails=1";
try
{
NominatimReverseGeocodeResponse? response =
await _httpClient.GetFromJsonAsync<NominatimReverseGeocodeResponse>(
url,
cancellationToken
);
if (response?.Address is null || response.Error is not null)
return null;
string addressLine1 = string.Join(
' ',
new[] { response.Address.HouseNumber, response.Address.Road }.Where(part =>
!string.IsNullOrWhiteSpace(part)
)
);
return string.IsNullOrWhiteSpace(addressLine1)
? null
: new ReverseGeocodeResult(addressLine1, response.Address.Postcode ?? string.Empty);
}
catch (HttpRequestException)
{
return null;
}
catch (TaskCanceledException) when (!cancellationToken.IsCancellationRequested)
{
// HttpClient's own timeout elapsed here.
return null;
}
}
private async Task WaitForRateLimitAsync(CancellationToken cancellationToken)
{
await _rateLimitLock.WaitAsync(cancellationToken);
try
{
TimeSpan elapsed = DateTime.UtcNow - _lastRequestUtc;
if (elapsed < MinRequestInterval)
await Task.Delay(MinRequestInterval - elapsed, cancellationToken);
_lastRequestUtc = DateTime.UtcNow;
}
finally
{
_rateLimitLock.Release();
}
}
public void Dispose()
{
_httpClient.Dispose();
_rateLimitLock.Dispose();
}
}

View File

@@ -7,4 +7,8 @@ public sealed record BreweryAddress
public required double Longitude { get; init; } public required double Longitude { get; init; }
public required double Latitude { get; init; } public required double Latitude { get; init; }
public required string AddressLine1 { get; init; }
public required string PostalCode { get; init; }
} }

View File

@@ -1,6 +1,5 @@
using Database.Connection.DependencyInjection; using Database.Connection.DependencyInjection;
using Database.Seed.DatabaseHelpers; using Database.Seed.DatabaseHelpers;
using Database.Seed.Geocoding;
using Database.Seed.PipelineData; using Database.Seed.PipelineData;
using Database.Seed.Sqlite; using Database.Seed.Sqlite;
using Domain.Entities; using Domain.Entities;
@@ -17,9 +16,6 @@ using Microsoft.Extensions.Configuration;
using Microsoft.Extensions.DependencyInjection; using Microsoft.Extensions.DependencyInjection;
using Spectre.Console; using Spectre.Console;
const string NominatimUserAgent =
"TheBiergartenApp-Seeder/1.0 (+https://github.com/aaronpo97/the-biergarten-app)";
return await RunAsync(); return await RunAsync();
static async Task<int> RunAsync() static async Task<int> RunAsync()
@@ -96,10 +92,6 @@ static async Task<int> RunAsync()
} }
); );
List<string> geocodingWarnings = [];
using (NominatimReverseGeocoder geocoder = new(NominatimUserAgent))
{
await AnsiConsole await AnsiConsole
.Status() .Status()
.Spinner(Spinner.Known.Dots) .Spinner(Spinner.Known.Dots)
@@ -111,19 +103,13 @@ static async Task<int> RunAsync()
await LoadBreweriesIntoDatabaseAsync( await LoadBreweriesIntoDatabaseAsync(
breweryRepository, breweryRepository,
locationRepository, locationRepository,
geocoder,
seedData.Breweries, seedData.Breweries,
postedByIds, postedByIds,
geocodingWarnings,
ctx ctx
); );
ctx.Status("Brewery data loaded into target database."); ctx.Status("Brewery data loaded into target database.");
} }
); );
}
foreach (string warning in geocodingWarnings)
AnsiConsole.MarkupLine($"[yellow]![/] {warning.EscapeMarkup()}");
return 0; return 0;
} }
@@ -175,15 +161,11 @@ static async Task<IReadOnlyList<Guid>> LoadUsersIntoDatabaseAsync(
static async Task LoadBreweriesIntoDatabaseAsync( static async Task LoadBreweriesIntoDatabaseAsync(
IBreweryRepository breweryRepository, IBreweryRepository breweryRepository,
ILocationRepository locationRepository, ILocationRepository locationRepository,
NominatimReverseGeocoder geocoder,
IReadOnlyList<BreweryRecord> breweries, IReadOnlyList<BreweryRecord> breweries,
IReadOnlyList<Guid> posterUserIds, IReadOnlyList<Guid> posterUserIds,
List<string> geocodingWarnings,
StatusContext ctx StatusContext ctx
) )
{ {
const string placeholderAddressLine1 = "Address unavailable";
if (posterUserIds.Count == 0) if (posterUserIds.Count == 0)
throw new InvalidOperationException("Cannot load breweries without any registered users."); throw new InvalidOperationException("Cannot load breweries without any registered users.");
@@ -191,9 +173,7 @@ static async Task LoadBreweriesIntoDatabaseAsync(
{ {
BreweryRecord breweryRecord = breweries[i]; BreweryRecord breweryRecord = breweries[i];
ctx.Status( ctx.Status($"Loading brewery {i + 1}/{breweries.Count} into target database...");
$"Geocoding and loading brewery {i + 1}/{breweries.Count} into target database..."
);
Guid cityId = await locationRepository.GetOrCreateCityIdAsync( Guid cityId = await locationRepository.GetOrCreateCityIdAsync(
new CityLocation( new CityLocation(
@@ -205,18 +185,6 @@ static async Task LoadBreweriesIntoDatabaseAsync(
) )
); );
ReverseGeocodeResult? geocoded = await geocoder.ReverseGeocodeAsync(
breweryRecord.Address.Longitude,
breweryRecord.Address.Latitude
);
if (geocoded is null)
geocodingWarnings.Add(
$"Could not reverse geocode '{breweryRecord.Brewery.NameEn}' "
+ $"({breweryRecord.Address.Latitude}, {breweryRecord.Address.Longitude}); "
+ "using a placeholder address."
);
await breweryRepository.CreateAsync( await breweryRepository.CreateAsync(
new BreweryPost new BreweryPost
{ {
@@ -229,8 +197,8 @@ static async Task LoadBreweriesIntoDatabaseAsync(
{ {
BreweryPostLocationId = Guid.NewGuid(), BreweryPostLocationId = Guid.NewGuid(),
CityId = cityId, CityId = cityId,
AddressLine1 = geocoded?.AddressLine1 ?? placeholderAddressLine1, AddressLine1 = breweryRecord.Address.AddressLine1,
PostalCode = geocoded?.PostalCode ?? string.Empty, PostalCode = breweryRecord.Address.PostalCode,
}, },
} }
); );

View File

@@ -54,7 +54,7 @@ public sealed class PipelineSeedDataReader
{ {
const string sql = """ const string sql = """
SELECT b.name_en, b.description_en, b.name_local, b.description_local, SELECT b.name_en, b.description_en, b.name_local, b.description_local,
ba.city_id, ba.longitude, ba.latitude ba.city_id, ba.longitude, ba.latitude, ba.address_line1, ba.postal_code
FROM breweries b FROM breweries b
JOIN brewery_addresses ba ON ba.brewery_id = b.id; JOIN brewery_addresses ba ON ba.brewery_id = b.id;
"""; """;
@@ -69,6 +69,8 @@ public sealed class PipelineSeedDataReader
const int cityIdIndex = 4; const int cityIdIndex = 4;
const int longitudeIndex = 5; const int longitudeIndex = 5;
const int latitudeIndex = 6; const int latitudeIndex = 6;
const int addressLine1Index = 7;
const int postalCodeIndex = 8;
List<BreweryRecord> records = []; List<BreweryRecord> records = [];
while (await reader.ReadAsync(cancellationToken)) while (await reader.ReadAsync(cancellationToken))
@@ -87,6 +89,8 @@ public sealed class PipelineSeedDataReader
City = cities[reader.GetInt64(cityIdIndex)], City = cities[reader.GetInt64(cityIdIndex)],
Longitude = reader.GetDouble(longitudeIndex), Longitude = reader.GetDouble(longitudeIndex),
Latitude = reader.GetDouble(latitudeIndex), Latitude = reader.GetDouble(latitudeIndex),
AddressLine1 = reader.GetString(addressLine1Index),
PostalCode = reader.GetString(postalCodeIndex),
}, },
} }
); );