Add support for ZSTD compression

[flavioheleno]

Change Log

Added

• zstd compression functions

Fixed

Changed

Removed

Deprecated

Security

---

Description

This PR adds support for ZSTD compression in Parquet Codec.

Closes #782.

[github-actions]

Flow PHP - Benchmarks

^{_{Results of the benchmarks from this PR are compared with the results from 1.x branch.}}

Extractors

+-----------------------+-------------------+------+-----+------------------+------------------+-----------------+
| benchmark             | subject           | revs | its | mem_peak         | mode             | rstdev          |
+-----------------------+-------------------+------+-----+------------------+------------------+-----------------+
| CSVExtractorBench     | bench_extract_10k | 1    | 3   | 3.910mb +0.04%   | 510.027ms +0.18% | ±2.28% +372.28% |
| JsonExtractorBench    | bench_extract_10k | 1    | 3   | 3.943mb +0.04%   | 1.071s +0.83%    | ±0.44% -32.66%  |
| ParquetExtractorBench | bench_extract_10k | 1    | 3   | 135.375mb +0.00% | 759.438ms +0.96% | ±0.60% -65.62%  |
| TextExtractorBench    | bench_extract_10k | 1    | 3   | 3.670mb +0.04%   | 33.799ms -0.66%  | ±0.81% +47.02%  |
| XmlExtractorBench     | bench_extract_10k | 1    | 3   | 3.617mb +0.04%   | 434.124ms -1.05% | ±0.35% -83.45%  |
+-----------------------+-------------------+------+-----+------------------+------------------+-----------------+

Transformers

+-----------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
| benchmark                   | subject                  | revs | its | mem_peak         | mode            | rstdev         |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+----------------+
| RenameEntryTransformerBench | bench_transform_10k_rows | 1    | 3   | 115.962mb +0.00% | 61.712ms +1.10% | ±0.28% -83.79% |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+----------------+

Loaders

+--------------------+----------------+------+-----+------------------+-----------------+-----------------+
| benchmark          | subject        | revs | its | mem_peak         | mode            | rstdev          |
+--------------------+----------------+------+-----+------------------+-----------------+-----------------+
| CSVLoaderBench     | bench_load_10k | 1    | 3   | 54.065mb +0.00%  | 86.840ms +0.06% | ±1.13% +17.75%  |
| JsonLoaderBench    | bench_load_10k | 1    | 3   | 106.498mb +0.00% | 53.481ms -1.31% | ±0.46% -29.82%  |
| ParquetLoaderBench | bench_load_10k | 1    | 3   | 225.831mb +0.00% | 1.430s -0.17%   | ±1.10% -0.78%   |
| TextLoaderBench    | bench_load_10k | 1    | 3   | 16.859mb +0.01%  | 44.167ms +0.33% | ±1.60% +349.74% |
+--------------------+----------------+------+-----+------------------+-----------------+-----------------+

Building Blocks

+-------------------------+----------------------------+------+-----+------------------+------------------+-------------------------------+
| benchmark               | subject                    | revs | its | mem_peak         | mode             | rstdev                        |
+-------------------------+----------------------------+------+-----+------------------+------------------+-------------------------------+
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 116.513mb +0.00% | 501.640ms +0.22% | ±0.55% -60.87%                |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 59.991mb +0.00%  | 253.276ms +1.80% | ±0.99% +66.05%                |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 14.924mb +0.01%  | 56.272ms +3.12%  | ±1.81% -41.09%                |
| TypeDetectorBench       | bench_type_detector        | 1    | 3   | 59.691mb +0.00%  | 466.663ms +7.69% | ±1.75% +86.05%                |
| TypeDetectorBench       | bench_type_detector        | 1    | 3   | 14.230mb +0.01%  | 88.526ms +2.12%  | ±0.35% -71.15%                |
| RowsBench               | bench_chunk_10_on_10k      | 2    | 3   | 86.782mb +0.00%  | 3.814ms +4.89%   | ±1.62% -41.38%                |
| RowsBench               | bench_diff_left_1k_on_10k  | 2    | 3   | 102.380mb +0.00% | 188.975ms -0.62% | ±0.87% +210.95%               |
| RowsBench               | bench_diff_right_1k_on_10k | 2    | 3   | 85.100mb +0.00%  | 19.513ms +1.37%  | ±1.49% +349.35%               |
| RowsBench               | bench_drop_1k_on_10k       | 2    | 3   | 88.022mb +0.00%  | 2.150ms +4.47%   | ±3.58% +3175.98%              |
| RowsBench               | bench_drop_right_1k_on_10k | 2    | 3   | 88.022mb +0.00%  | 1.900ms -2.32%   | ±0.93% -74.61%                |
| RowsBench               | bench_entries_on_10k       | 2    | 3   | 85.134mb +0.00%  | 2.945ms -0.65%   | ±3.66% +331.95%               |
| RowsBench               | bench_filter_on_10k        | 2    | 3   | 85.663mb +0.00%  | 15.247ms -9.91%  | ±2.72% +348.49%               |
| RowsBench               | bench_find_on_10k          | 2    | 3   | 85.663mb +0.00%  | 15.312ms -7.97%  | ±1.56% +30.05%                |
| RowsBench               | bench_find_one_on_10k      | 10   | 3   | 83.568mb +0.00%  | 1.906μs +0.32%   | ±2.44% +20864134789308000.00% |
| RowsBench               | bench_first_on_10k         | 10   | 3   | 83.568mb +0.00%  | 0.400μs 0.00%    | ±0.00% 0.00%                  |
| RowsBench               | bench_flat_map_on_1k       | 2    | 3   | 92.917mb +0.00%  | 12.751ms +0.75%  | ±3.69% +92.95%                |
| RowsBench               | bench_map_on_10k           | 2    | 3   | 122.288mb +0.00% | 63.123ms +0.31%  | ±0.96% +43.92%                |
| RowsBench               | bench_merge_1k_on_10k      | 2    | 3   | 86.183mb +0.00%  | 1.784ms +10.30%  | ±1.29% -42.89%                |
| RowsBench               | bench_partition_by_on_10k  | 2    | 3   | 89.530mb +0.00%  | 65.046ms +1.33%  | ±0.68% -9.30%                 |
| RowsBench               | bench_remove_on_10k        | 2    | 3   | 88.284mb +0.00%  | 4.203ms -4.59%   | ±1.31% -56.75%                |
| RowsBench               | bench_sort_asc_on_1k       | 2    | 3   | 83.645mb +0.08%  | 39.955ms +0.18%  | ±1.62% +3.54%                 |
| RowsBench               | bench_sort_by_on_1k        | 2    | 3   | 83.646mb +0.08%  | 41.468ms +0.58%  | ±1.56% -12.87%                |
| RowsBench               | bench_sort_desc_on_1k      | 2    | 3   | 83.645mb +0.08%  | 40.520ms +2.26%  | ±1.96% -31.32%                |
| RowsBench               | bench_sort_entries_on_1k   | 2    | 3   | 86.009mb +0.00%  | 7.525ms -5.29%   | ±3.49% +47.20%                |
| RowsBench               | bench_sort_on_1k           | 2    | 3   | 83.568mb +0.00%  | 29.146ms +1.85%  | ±0.92% -15.66%                |
| RowsBench               | bench_take_1k_on_10k       | 10   | 3   | 83.568mb +0.00%  | 14.817μs +7.21%  | ±1.47% +62.21%                |
| RowsBench               | bench_take_right_1k_on_10k | 10   | 3   | 83.568mb +0.00%  | 17.699μs +6.14%  | ±1.38% -55.27%                |
| RowsBench               | bench_unique_on_1k         | 2    | 3   | 102.382mb +0.00% | 195.093ms +0.56% | ±1.47% +382.01%               |
+-------------------------+----------------------------+------+-----+------------------+------------------+-------------------------------+

[norberttech]

that was quick, thank you @flavioheleno 🍻