Conversation
- v0.1.2 -> v.0.1.3
Adds ChunkedBackend<T>, a StorageBackend that partitions values into fixed-size Vec<Vec<T>> chunks. A new chunk is allocated only when the current one is full, so no existing data is copied on append — the main advantage over ColumnarBackend for high-frequency streaming workloads. - crates/storage/chunked.rs: full implementation with new(), chunk_size(), n_chunks(), and the complete StorageBackend<T> trait (len, is_empty, get, push, slice, iter). get() is O(1) via chunk/offset index arithmetic. - crates/storage/mod.rs: export ChunkedBackend alongside existing backends. - crates/storage/backend.rs: add ChunkedBackend to the trait doc table. - tests/storage/test_chunked.rs: 20 unit tests covering construction (exact/partial/oversized/empty chunks), get (first, last, chunk boundary, out-of-bounds), push (triggers new chunk, fills partial, works on empty backend), slice (subset, preserved chunk_size, source unchanged), and iter. - examples/temporal_series_with_chunked_backend.rs: demonstrates construction from existing data, get, iteration, incremental appends, and slice. - README.md: Storage Backends section extended with ChunkedBackend subsection, comparison table updated, new example added to the examples table. - ROADMAP.md: mark ChunkedBackend item as complete. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds four new sliding-window methods to RollingSeries: - sum: window sum with NaN fill during warm-up - min/max: window extremes via fold over INFINITY / NEG_INFINITY - std: Bessel-corrected sample std deviation; rejects window < 2 with InvalidWindow Includes 25 new tests (given/when/then) across test_sum, test_min, test_max, and test_std, and four new Criterion benchmarks (window=20, n=1000) alongside the existing rolling_mean benchmark. bench helper refactored to make_ts. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds read_panel_csv and write_panel_csv to crates/io/csv.rs. The format is one symbol per column with a shared index column: index,AAPL,MSFT 1,150.0,300.0 2,152.0,305.0 read_panel_csv parses the header to extract symbol names, validates field counts per row, and delegates construction to Panel::new so all existing invariants are enforced. write_panel_csv serialises in the same layout; the round-trip is value-lossless for all f64 values with an exact decimal representation. Also adds: - Panel::index() accessor (required by the writer to iterate timestamps) - 14 integration tests across test_read_panel_csv and test_write_panel_csv (given/when/then convention): shape, symbols, values, index preservation, non-existent file, single-symbol, malformed value, empty file, header-only, wrong field count, invalid path, round-trip, header format, row count - examples/panel_io.rs demonstrating the full round-trip - examples/panel_input.csv fixture committed for read tests - .gitignore entries for runtime-generated test CSVs Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds three missing-value imputation methods to TimeSeries under a new NAN HANDLING section in crates/series/time_series.rs: - forward_fill: carries the last valid value forward (LOCF); leading NaNs before the first valid observation are preserved - backward_fill: carries the next valid value backward (NOCB); trailing NaNs after the last valid observation are preserved - fill_with(fill_value): replaces every NaN with a fixed scalar; passing f64::NAN is a valid no-op All three return Result<Self, TemporalSeriesError> and preserve the original index unchanged, consistent with the rest of the series API. Includes 18 integration tests across test_forward_fill, test_backward_fill, and test_fill_with (6 each, given/when/then convention), covering interior gaps, leading/trailing NaN edge cases, all-NaN series, no-NaN identity, and output length + index preservation. Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Adds an optional `serde` feature flag that derives Serialize and
Deserialize for both TimeSeries and Panel:
cargo add temporalseries --features serde
The derives use the full-path form `serde::Serialize` / `serde::Deserialize`
so no unconditional serde import is needed in library code. Panel's private
fields are accessible to the derive macro because it generates code in the
same module scope.
JSON layout produced by serde_json:
TimeSeries → {"index":[...],"values":[...]}
Panel → {"index":[...],"symbols":[...],"values":[[...],...]}
NaN behaviour: serde_json maps NaN to null (lossy); formats that preserve
IEEE 754 special values (MessagePack, bincode) round-trip NaN cleanly.
Panel deserialization bypasses Panel::new — invariants hold for any Panel
serialized from a valid instance; untrusted JSON is not validated.
Also adds:
- serde_json dev-dependency for tests and the example
- 12 integration tests (6 per type, given/when/then) under tests/serde_support/,
gated via #[cfg(feature = "serde")] in tests/integration.rs: field names in
JSON, index / values / symbols / shape round-trip, raw JSON deserialization,
NaN → null behaviour
- examples/serde_support.rs (required-features = ["serde"]) with full
round-trip demo, hand-crafted JSON parsing, and NaN/null explanation
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.