Add support for BROTLI compression

[flavioheleno]

Change Log

Added

• Added support for BROTLI compression algorithm to parquet

Fixed

Changed

Removed

Deprecated

Security

---

Description

Add support for BROTLI compression.

Closes #784.

[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   | 3.913mb +0.04%   | 505.119ms -1.04% | ±0.37% +3.36%   |
| JsonExtractorBench    | bench_extract_10k | 1    | 3   | 3.946mb +0.04%   | 1.065s -0.86%    | ±0.26% -64.66%  |
| ParquetExtractorBench | bench_extract_10k | 1    | 3   | 135.382mb +0.00% | 734.997ms -1.12% | ±0.48% -57.48%  |
| TextExtractorBench    | bench_extract_10k | 1    | 3   | 3.672mb +0.05%   | 34.205ms +1.91%  | ±1.34% +236.46% |
| XmlExtractorBench     | bench_extract_10k | 1    | 3   | 3.619mb +0.05%   | 429.989ms -1.13% | ±0.70% -51.56%  |
+-----------------------+-------------------+------+-----+------------------+------------------+-----------------+

Transformers

+-----------------------------+--------------------------+------+-----+------------------+-----------------+-----------------+
| benchmark                   | subject                  | revs | its | mem_peak         | mode            | rstdev          |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+-----------------+
| RenameEntryTransformerBench | bench_transform_10k_rows | 1    | 3   | 115.965mb +0.00% | 59.003ms -2.15% | ±1.32% +118.81% |
+-----------------------------+--------------------------+------+-----+------------------+-----------------+-----------------+

Loaders

+--------------------+----------------+------+-----+------------------+-----------------+------------------+
| benchmark          | subject        | revs | its | mem_peak         | mode            | rstdev           |
+--------------------+----------------+------+-----+------------------+-----------------+------------------+
| CSVLoaderBench     | bench_load_10k | 1    | 3   | 54.068mb +0.00%  | 85.177ms +0.97% | ±0.78% -54.69%   |
| JsonLoaderBench    | bench_load_10k | 1    | 3   | 106.500mb +0.00% | 51.186ms -3.71% | ±0.25% -77.08%   |
| ParquetLoaderBench | bench_load_10k | 1    | 3   | 225.839mb +0.00% | 1.390s -0.17%   | ±0.16% -78.98%   |
| TextLoaderBench    | bench_load_10k | 1    | 3   | 16.862mb +0.01%  | 43.578ms -1.01% | ±1.25% +1124.97% |
+--------------------+----------------+------+-----+------------------+-----------------+------------------+

Building Blocks

+-------------------------+----------------------------+------+-----+------------------+------------------+------------------+
| benchmark               | subject                    | revs | its | mem_peak         | mode             | rstdev           |
+-------------------------+----------------------------+------+-----+------------------+------------------+------------------+
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 116.515mb +0.00% | 493.866ms -1.95% | ±1.97% -30.46%   |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 59.993mb +0.00%  | 249.077ms -1.65% | ±0.58% -37.44%   |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 14.927mb +0.01%  | 52.951ms +0.32%  | ±0.39% -55.51%   |
| TypeDetectorBench       | bench_type_detector        | 1    | 3   | 59.694mb +0.00%  | 430.109ms -4.46% | ±0.27% -88.73%   |
| TypeDetectorBench       | bench_type_detector        | 1    | 3   | 14.233mb +0.01%  | 84.435ms -1.58%  | ±0.33% -70.19%   |
| RowsBench               | bench_chunk_10_on_10k      | 2    | 3   | 86.785mb +0.00%  | 3.222ms -6.19%   | ±1.13% -68.83%   |
| RowsBench               | bench_diff_left_1k_on_10k  | 2    | 3   | 102.383mb +0.00% | 192.301ms +2.13% | ±0.30% -87.29%   |
| RowsBench               | bench_diff_right_1k_on_10k | 2    | 3   | 85.103mb +0.00%  | 18.958ms -0.73%  | ±0.48% +116.29%  |
| RowsBench               | bench_drop_1k_on_10k       | 2    | 3   | 88.025mb +0.00%  | 1.755ms +2.59%   | ±1.52% -37.05%   |
| RowsBench               | bench_drop_right_1k_on_10k | 2    | 3   | 88.025mb +0.00%  | 1.684ms -10.92%  | ±1.79% +648.50%  |
| RowsBench               | bench_entries_on_10k       | 2    | 3   | 85.137mb +0.00%  | 2.636ms +3.23%   | ±1.94% +33.68%   |
| RowsBench               | bench_filter_on_10k        | 2    | 3   | 85.666mb +0.00%  | 14.921ms +2.40%  | ±1.37% +1205.70% |
| RowsBench               | bench_find_on_10k          | 2    | 3   | 85.666mb +0.00%  | 14.925ms +1.77%  | ±1.03% +31.07%   |
| RowsBench               | bench_find_one_on_10k      | 10   | 3   | 83.570mb +0.00%  | 1.694μs +5.47%   | ±2.83% -2.00%    |
| RowsBench               | bench_first_on_10k         | 10   | 3   | 83.570mb +0.00%  | 0.300μs +50.00%  | ±0.00% -100.00%  |
| RowsBench               | bench_flat_map_on_1k       | 2    | 3   | 92.920mb +0.00%  | 12.158ms +0.42%  | ±1.02% +656.47%  |
| RowsBench               | bench_map_on_10k           | 2    | 3   | 122.291mb +0.00% | 60.177ms -2.29%  | ±0.54% +15.38%   |
| RowsBench               | bench_merge_1k_on_10k      | 2    | 3   | 86.186mb +0.00%  | 1.268ms +3.07%   | ±3.38% +121.77%  |
| RowsBench               | bench_partition_by_on_10k  | 2    | 3   | 89.533mb +0.00%  | 64.409ms +5.37%  | ±1.56% -48.84%   |
| RowsBench               | bench_remove_on_10k        | 2    | 3   | 88.287mb +0.00%  | 3.834ms -0.93%   | ±0.45% +234.01%  |
| RowsBench               | bench_sort_asc_on_1k       | 2    | 3   | 83.714mb +0.00%  | 39.006ms -0.35%  | ±0.54% -39.22%   |
| RowsBench               | bench_sort_by_on_1k        | 2    | 3   | 83.714mb +0.00%  | 39.849ms +1.68%  | ±0.40% -61.07%   |
| RowsBench               | bench_sort_desc_on_1k      | 2    | 3   | 83.714mb +0.00%  | 39.047ms -1.94%  | ±0.37% -79.11%   |
| RowsBench               | bench_sort_entries_on_1k   | 2    | 3   | 86.011mb +0.00%  | 7.383ms +0.48%   | ±3.47% +363.94%  |
| RowsBench               | bench_sort_on_1k           | 2    | 3   | 83.570mb +0.00%  | 28.559ms +0.64%  | ±0.86% +611.96%  |
| RowsBench               | bench_take_1k_on_10k       | 10   | 3   | 83.570mb +0.00%  | 13.300μs -0.22%  | ±1.23% -29.85%   |
| RowsBench               | bench_take_right_1k_on_10k | 10   | 3   | 83.570mb +0.00%  | 16.234μs +4.65%  | ±1.27% +109.86%  |
| RowsBench               | bench_unique_on_1k         | 2    | 3   | 102.384mb +0.00% | 193.549ms +0.22% | ±0.27% -18.34%   |
+-------------------------+----------------------------+------+-----+------------------+------------------+------------------+

[norberttech]

Before I merge this one (once conflicts are resolved) would you mind to also mention extensions required to use particular compression algorithms in parquet docs? At the bottom of this page

[flavioheleno]

for sure!