public async Task LoadNoReplace_RecordExists_ReturnsOldRecord()
{
// ARRANGE
var e = CosmosDBSqlApiArranger.GetEtlEventMock("EtlEvent_2018-05-22T01:00:00.000000Z");
var datetime = DateTime.UtcNow;
DocumentLoader sut = new DocumentLoader(
client,
"cafdb",
"items");
// ACT
ResourceResponse <Document> result = await sut.LoadNoReplace(e);
// ASSERT
Assert.True(result.StatusCode == HttpStatusCode.OK);
Assert.True(result.Resource.Timestamp < datetime);
}
public async Task LoadNoReplace_RecordDoesNotExist_ReturnsNewRecord()
{
// ARRANGE
var datetime = DateTime.UtcNow;
DocumentClient client = new DocumentClient(
new Uri("https://localhost:8081"),
"C2y6yDjf5/R+ob0N8A7Cgv30VRDJIWEHLM+4QDU5DE2nQ9nDuVTqobD4b8mGGyPMbIZnqyMsEcaGQy67XIw/Jw==");
var e = CosmosDBSqlApiArranger.GetEtlEventMock($"EtlEvent_{DateTime.UtcNow.ToString("o")}");
DocumentLoader sut = new DocumentLoader(
client,
"cafdb",
"items");
// ACT
ResourceResponse <Document> result = await sut.LoadNoReplace(e);
// ASSERT
Assert.True(result.StatusCode == HttpStatusCode.Created);
Assert.True(result.Resource.Timestamp > datetime);
}
public async Task LoadNoReplace_RecordExists_ReturnsOldRecord()
{
// ARRANGE
DocumentClient client = new DocumentClient(
new Uri("https://localhost:8081"),
"C2y6yDjf5/R+ob0N8A7Cgv30VRDJIWEHLM+4QDU5DE2nQ9nDuVTqobD4b8mGGyPMbIZnqyMsEcaGQy67XIw/Jw==");
var e = CosmosDBSqlApiArranger.GetEtlEventMock("EtlEvent_2018-05-22T01:00:00.000000Z");
var datetime = DateTime.UtcNow;
DocumentLoader sut = new DocumentLoader(
client,
"cafdb",
"items");
// ACT
ResourceResponse <Document> result = await sut.LoadNoReplace(e);
// ASSERT
Assert.True(result.StatusCode == HttpStatusCode.OK);
Assert.True(result.Resource.Timestamp < datetime);
}
public async Task LoadNoReplace_RecordDoesNotExist_ReturnsNewRecord()
{
// ARRANGE
var datetime = DateTime.UtcNow;
var e = CosmosDBSqlApiArranger.GetEtlEventMock($"EtlEvent_{DateTime.UtcNow.ToString("o")}");
DocumentLoader sut = new DocumentLoader(
client,
"cafdb",
"items");
// ACT
ResourceResponse <Document> result = await sut.LoadNoReplace(e);
// ASSERT
Assert.True(result.StatusCode == HttpStatusCode.Created);
Assert.InRange <DateTime>(
result.Resource.Timestamp,
datetime.Add(new TimeSpan(0, -1, 0)),
datetime.Add(new TimeSpan(0, 1, 0)));
}
public static async Task Run(
[BlobTrigger("ectower-cookeast/raw/Flux/{name}", Connection = "ltarcafdatastreamConnectionString")] Stream myBlob,
string name,
TraceWriter log,
ExecutionContext context)
{
log.Info($"C# Blob trigger function Processed blob\n Name:{name} \n Size: {myBlob.Length} Bytes");
//var config = new ConfigurationBuilder()
// .SetBasePath(context.FunctionAppDirectory)
// .AddJsonFile("local.settings.json", optional: true, reloadOnChange: true)
// .AddEnvironmentVariables()
// .Build();
EtlEvent etlEvent = new EtlEvent(
"EtlEvent",
"AzureFunction",
"http://files.cafltar.org/data/schema/documentDb/v2/etlEvent.json",
"CafMeteorologyEcTower",
"1.1", "LoggerNetFluxToCosmosDBSqlApiMeasurementCookEast",
DateTime.UtcNow);
etlEvent.Outputs = null;
etlEvent.Inputs.Add($"ectower-cookeast/raw/Flux/{name}");
etlEvent.Logs.Add($"C# Blob trigger function Processed blob\n Name:{name} \n Size: {myBlob.Length} Bytes");
StreamReader reader = new StreamReader(myBlob);
string contents = "";
log.Info("About to read contents");
try
{
contents = reader.ReadToEnd();
}
catch (Exception e)
{
etlEvent.Logs.Add(
$"Error reading Blob: {e.Message}");
}
//DocumentClient client = new DocumentClient(
// new Uri(
// config["Values:AzureCosmosDBUri"]),
// config["Values:AzureCosmosDBKey"]);
DocumentClient client;
try
{
client = new DocumentClient(
new Uri(
ConfigurationManager.AppSettings["AzureCosmosDBUri"]),
ConfigurationManager.AppSettings["AzureCosmosDBKey"]);
}
catch (Exception e)
{
etlEvent.Logs.Add(
$"Error creating DocumentClient: {e.Message}");
log.Error($"Error creating DocumentClient: {e.Message}");
throw new Exception("Error creating DocumentClient", e);
}
DocumentLoader loader = new DocumentLoader(
client,
"cafdb",
"items");
log.Info("Created client and loader");
if (!String.IsNullOrEmpty(contents))
{
try
{
log.Info("Attempting extract and transform");
TOA5Extractor extractor = new TOA5Extractor(
name,
contents,
-8);
TOA5 fluxTable = extractor.GetTOA5 <Flux>();
// TODO: Move strings and such to settings file
DocumentDbMeasurementV2Transformer transformer =
new DocumentDbMeasurementV2Transformer(
new MapFromFluxDataTableToCafStandards(),
"http://files.cafltar.org/data/schema/documentDb/v2/measurement.json",
etlEvent.Id,
"Measurement",
"CafMeteorologyEcTower",
1800);
List <MeasurementV2> measurements =
transformer.ToMeasurements(fluxTable);
log.Info("Attempting load");
/// Using the bulkImport sproc doesn't provide much benefit since
/// most data tables will only have a few measurements with the
/// same partition key. But it's better than nothing.
StoredProcedureResponse <bool>[] results = await loader.LoadBulk(measurements);
log.Info($"Loaded {results.Length.ToString()} measurements");
}
catch (Exception e)
{
etlEvent.Logs.Add(
$"Error in ETL pipeline: {e.Message}");
log.Error($"Error in ETL pipeline: {e.Message}");
throw new Exception("Error in ETL pipeline", e);
}
finally
{
log.Info("Loading etlEvent to db");
etlEvent.DateTimeEnd = DateTime.UtcNow;
ResourceResponse <Document> result = await loader.LoadNoReplace(etlEvent);
log.Info($"Result of writing EtlEvent: {result.StatusCode.ToString()}");
}
log.Info("Function completed");
}
}
public async Task PipeItAsync()
{
EtlEvent etlEvent = new EtlEvent(
"EtlEvent",
"AzureFunction",
"http://files.cafltar.org/data/schema/documentDb/v2/etlEvent.json",
"CafMeteorologyEcTower",
version, functionName,
DateTime.UtcNow);
etlEvent.Inputs.Add(blobPath);
StreamReader reader = new StreamReader(myBlob);
string contents = "";
log.LogInformation("About to read contents");
try
{
contents = reader.ReadToEnd();
}
catch (Exception e)
{
etlEvent.Logs.Add(
$"Error reading Blob: {e.Message}");
}
DocumentLoader loader = new DocumentLoader(
client,
"cafdb",
"items");
log.LogInformation("Created loader");
if (!String.IsNullOrEmpty(contents))
{
try
{
log.LogInformation("Attempting extract and transform");
TOA5Extractor extractor = new TOA5Extractor(
name,
contents,
-8);
TOA5 toa5 = extractor.GetTOA5(observation);
CosmosDBSqlApiV2Transformer transformer =
new CosmosDBSqlApiV2Transformer(
new MapFromToa5DataTableToCafStandards(),
"http://files.cafltar.org/data/schema/documentDb/v2/measurement.json",
etlEvent.Id,
"Measurement",
"CafMeteorologyEcTower",
timestep);
List <MeasurementV2> measurements =
transformer.ToMeasurements(toa5);
log.LogInformation("Attempting load");
int docsLoaded = 0;
int docsError = 0;
foreach (MeasurementV2 measurement in measurements)
{
try
{
ResourceResponse <Document> result =
await loader.LoadNoReplace(measurement);
if (
result.StatusCode == HttpStatusCode.Created ||
result.StatusCode == HttpStatusCode.OK)
{
etlEvent.Outputs.Add(result.Resource.Id);
docsLoaded++;
}
else
{
etlEvent.Logs.Add(
$"StatusCode: {result.StatusCode} on MeasurementV2: {measurement.Id.ToString()}");
docsError++;
}
}
catch (Exception e)
{
etlEvent.Logs.Add(
$"Error loading {measurement.Id.ToString()} MeasurementV2: {e.Message}");
log.LogError($"Error loading MeasurementV2: {e.Message}");
docsError++;
}
}
log.LogInformation(
$"Loaded {docsLoaded.ToString()} MeasurementV2s.");
log.LogInformation(
$"Error loading {docsError.ToString()} MeasurementV2s.");
etlEvent.Logs.Add(
$"Loaded {docsLoaded.ToString()} MeasurementV2s");
etlEvent.Logs.Add(
$"Error loading {docsError.ToString()} MeasurementV2s");
}
catch (Exception e)
{
etlEvent.Logs.Add(
$"Error in ETL pipeline: {e.Message}");
log.LogError($"Error in ETL pipeline: {e.Message}");
throw new Exception("Error in ETL pipeline", e);
}
finally
{
log.LogInformation("Loading etlEvent to db");
etlEvent.DateTimeEnd = DateTime.UtcNow;
ResourceResponse <Document> result = await loader.LoadNoReplace(etlEvent);
log.LogInformation($"Result of writing EtlEvent: {result.StatusCode.ToString()}");
}
}
}
static async Task MainAsync()
{
EtlEvent etlEvent = new EtlEvent(
"EtlEvent",
"LocalProcess",
"http://files.cafltar.org/data/schema/documentDb/v2/etlEvent.json",
"CookEastSoilGridPointSurvey",
"1.0",
"CookEastSoilGridPointSurvey_DotNet_SoilGridPointToCosmosDB",
DateTime.UtcNow);
var builder = new ConfigurationBuilder()
.SetBasePath(Directory.GetCurrentDirectory())
.AddJsonFile("appsettings.json");
var configuration = builder.Build();
JsonSerializerSettings serializerSettings = new JsonSerializerSettings
{
NullValueHandling = NullValueHandling.Ignore
};
string data = configuration["PathToData"];
string dict = configuration["PathToDictionary"];
if (!File.Exists(data) |
!File.Exists(dict))
{
throw new FileNotFoundException();
}
etlEvent.Inputs.Add(data);
etlEvent.Inputs.Add(dict);
DocumentClient client;
try
{
client = new DocumentClient(
new Uri(
configuration["CosmosServiceEndpoint"]),
configuration["CosmosAuthKey"],
serializerSettings);
}
catch (Exception e)
{
etlEvent.Logs.Add(
$"Error creating DocumentClient: {e.Message}");
throw new Exception("Error creating DocumentClient", e);
}
var extractor = new TidyDataCsvExtractor(
configuration["PathToData"],
configuration["PathToDictionary"]);
var transformer = new CosmosDBSqlApiSampleV2Transformer
<SoilGridPointSurveyV1, SoilSample>(
new MapFromSoilGridPointSurveyToSoilSample(),
"http://files.cafltar.org/data/schema/documentDb/v2/sample.json",
etlEvent.Id,
"CookEastSoilGridPointSurvey",
"CookEast",
"SoilSample");
var loader = new DocumentLoader(
client,
"cafdb",
"items");
try
{
TidyData extracted = extractor.Extract <SoilGridPointSurveyV1>();
etlEvent.Logs.Add(
$"Extracted TidyData with {extracted.Observations.Count} observations");
List <SoilSample> transformed = transformer.Transform(extracted);
etlEvent.Logs.Add(
$"Transformed TidyData to {transformed.Count} SoilSamples");
int docsLoaded = 0;
int docsError = 0;
foreach (SoilSample sample in transformed)
{
ResourceResponse <Document> result =
await loader.LoadNoReplace(sample);
if (result.StatusCode == HttpStatusCode.Created)
{
etlEvent.Outputs.Add(result.Resource.Id);
docsLoaded++;
}
else
{
docsError++;
}
// Notify data written then sleep to conserve Cosmos RU
Console.Write(".");
Thread.Sleep(40);
}
etlEvent.Logs.Add(
$"Loaded {docsLoaded.ToString()} SoilSamples");
etlEvent.Logs.Add(
$"Error loading {docsError.ToString()} SoilSamples");
}
catch (Exception e)
{
etlEvent.Logs.Add(
$"Error in ETL pipeline: {e.Message}");
throw new Exception("Error in ETL pipeline", e);
}
finally
{
etlEvent.DateTimeEnd = DateTime.UtcNow;
ResourceResponse <Document> result = await loader.LoadNoReplace(etlEvent);
}
}