blob: bf4cd9bfb38bbb4081696c33a815375c6f1b495d [file] [log] [blame]
// Copyright 2020 The Fuchsia Authors. All rights reserved.
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#include "src/local_aggregation_1_1/aggregation_procedures/string_histogram_aggregation_procedure.h"
#include <tuple>
#include "src/lib/util/datetime_util.h"
#include "src/lib/util/hash.h"
#include "src/local_aggregation_1_1/aggregation_procedures/aggregation_procedure.h"
#include "src/logger/encoder.h"
#include "src/pb/observation.pb.h"
#include "src/public/lib/statusor/status_macros.h"
namespace cobalt::local_aggregation {
void StringHistogramAggregationProcedure::UpdateAggregateData(
const logger::EventRecord &event_record, AggregateData *aggregate_data,
AggregationPeriodBucket *bucket) {
StringHistogram *histogram = aggregate_data->mutable_string_histogram();
std::string bytes =
util::FarmhashFingerprint(event_record.event()->string_event().string_value());
for (int i = 0; i < bucket->string_hashes_size(); i++) {
if (bucket->string_hashes(i) == bytes) {
(*histogram->mutable_histogram())[i] += 1;
return;
}
}
if (bucket->string_hashes_size() < string_buffer_max_) {
// Add new entry
(*histogram->mutable_histogram())[bucket->string_hashes_size()] += 1;
bucket->add_string_hashes(bytes);
}
}
lib::statusor::StatusOr<std::unique_ptr<Observation>>
StringHistogramAggregationProcedure::GenerateHourlyObservation(
const AggregateDataToGenerate &bucket) {
std::vector<std::tuple<std::vector<uint32_t>, std::vector<std::tuple<uint32_t, int64_t>>>> data;
data.reserve(bucket.aggregate_data.size());
for (const EventCodesAggregateData *aggregate_data : bucket.aggregate_data) {
const StringHistogram &histogram = aggregate_data->data().string_histogram();
std::vector<uint32_t> event_codes(aggregate_data->event_codes().begin(),
aggregate_data->event_codes().end());
std::vector<std::tuple<uint32_t, int64_t>> event_code_histogram;
for (const auto &[index, count] : histogram.histogram()) {
event_code_histogram.emplace_back(index, count);
}
data.emplace_back(event_codes, event_code_histogram);
}
if (data.empty()) {
return {nullptr};
}
std::vector<std::string> hashes;
for (const std::string &hash : bucket.string_hashes) {
hashes.push_back(hash);
}
return logger::Encoder::EncodeStringHistogramObservation(hashes, data);
}
std::string StringHistogramAggregationProcedure::DebugString() const { return "STRING_HISTOGRAM"; }
} // namespace cobalt::local_aggregation