DFTracer Views & Predicates

Namespace: dftracer::utils::utilities::composites::dft::views

struct ViewBuilderInput

Public Functions

ViewBuilderInput &with_view(const ViewDefinition &v)
ViewBuilderInput &with_file_path(const std::string &path)
ViewBuilderInput &with_index_path(const std::string &path)
ViewBuilderInput &with_uncompressed_size(std::size_t s)
ViewBuilderInput &with_num_checkpoints(std::size_t n)
ViewBuilderInput &with_bloom_cache(indexing::BloomFilterCache *c)
ViewBuilderInput &with_time_range(double begin, double end)

Public Members

ViewDefinition view
std::string file_path
std::string index_path
std::size_t uncompressed_size = 0
std::size_t num_checkpoints = 0
indexing::BloomFilterCache *bloom_cache = nullptr
std::optional<std::pair<double, double>> time_range
struct ViewBuilderOutput

Public Members

bool file_may_match = false
std::vector<ViewChunkCandidate> candidates
std::uint64_t total_checkpoints = 0
std::uint64_t skipped_checkpoints = 0
bool success = false
class ViewBuilderUtility : public dftracer::utils::utilities::Utility<ViewBuilderInput, ViewBuilderOutput, tags::Parallelizable>

Public Functions

coro::CoroTask<ViewBuilderOutput> process(const ViewBuilderInput &input) override
struct ViewChunkCandidate

Public Members

std::uint64_t checkpoint_idx = 0
std::size_t start_byte = 0
std::size_t end_byte = 0
struct ViewDefinition

Named view definition with optional query filter.

Public Functions

ViewDefinition &with_name(const std::string &n)
ViewDefinition &with_description(const std::string &d)
ViewDefinition &with_query(const std::string &query_str)

Set query from a DSL string. Silently ignored if parse fails.

ViewDefinition &with_query(Query q)
ViewDefinition &with_include_metadata(bool v)
std::string to_json() const

Public Members

std::string name

View name.

std::string description

Human-readable description.

std::optional<Query> query

Event filter (nullopt = match all).

bool include_metadata = true

Include ph=M metadata events.

Public Static Functions

static ViewDefinition from_json(const std::string &json)
static ViewDefinition io_view()

POSIX/STDIO I/O operations.

static ViewDefinition compute_view()

AI/HPC compute and framework operations.

static ViewDefinition dlio_view()

DLIO benchmark operations.

struct ViewReaderBatch

Public Functions

common::arrow::ArrowExportResult to_arrow() const
common::arrow::ArrowExportResult to_arrow(common::arrow::RecordBatchBuilder &builder) const

Public Members

std::vector<std::string_view> events

Event lines. In stream mode these are string_view into the decompressed chunk (zero copy, valid until next generator resume). Metadata events use owned strings stored in owned_events.

std::deque<std::string> owned_events

Owned storage for metadata events that outlive their source chunk. Uses deque so push_back doesn’t invalidate string_view refs.

std::uint64_t events_matched = 0
std::uint64_t events_scanned = 0
struct ViewReaderInput

Public Functions

ViewReaderInput &with_file_path(const std::string &path)
ViewReaderInput &with_index_path(const std::string &path)
ViewReaderInput &with_checkpoint_size(std::size_t sz)
ViewReaderInput &with_byte_range(std::size_t start, std::size_t end)
ViewReaderInput &with_checkpoint_idx(std::uint64_t idx)
ViewReaderInput &with_batch_size(std::size_t sz)
ViewReaderInput &with_event_batch_size(std::size_t sz)
ViewReaderInput &with_view(const ViewDefinition &v)

Public Members

std::string file_path
std::string index_path
std::size_t checkpoint_size = utilities::indexer::internal::Indexer::DEFAULT_CHECKPOINT_SIZE
std::size_t start_byte = 0
std::size_t end_byte = 0
std::uint64_t checkpoint_idx = 0
std::size_t batch_size = 4 * 1024 * 1024
std::size_t event_batch_size = 10000
ViewDefinition view
std::optional<common::query::Query> query
class ViewReaderUtility : public dftracer::utils::utilities::StreamingUtility<ViewReaderInput, ViewReaderBatch, tags::Parallelizable>

Public Functions

coro::AsyncGenerator<ViewReaderBatch> process(const ViewReaderInput &input) override