Skip to content

Added BatchSizeOptimization #788

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom
Merged
norberttech merged 4 commits into from
Nov 12, 2023
Merged

Added BatchSizeOptimization #788

norberttech merged 4 commits into from
Nov 12, 2023

Conversation

@norberttech
Copy link
Member

@norberttech norberttech commented Nov 12, 2023 •
edited
Loading

Change Log

Added

  • BatchSizeOptimization
  • httpClient option to Meilisearch loader configuration

Fixed

Changed

  • Optimizer can be now configured through ConfigBuilder
  • Execution Plan Processors can be now configured through ConfigBuilder

Removed

Deprecated

Security

Description

Closes: #738

@norberttech
Copy link
Member Author

norberttech commented Nov 12, 2023

@stloyd I noticed that Meilisearch has two methods, bulk_index, and bulk_update, but the update does not seem to be used in any way. Do you think it should be removed?

@github-actions
Copy link
Contributor

github-actions bot commented Nov 12, 2023 •
edited
Loading

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         |
+-----------------------+-------------------+------+-----+------------------+------------------+----------------+
| AvroExtractorBench    | bench_extract_10k | 1    | 3   | 35.274mb -0.03%  | 444.207ms +0.18% | ±0.09% -64.26% |
| CSVExtractorBench     | bench_extract_10k | 1    | 3   | 4.838mb +1.63%   | 340.241ms +0.22% | ±0.40% -3.20%  |
| JsonExtractorBench    | bench_extract_10k | 1    | 3   | 5.008mb +0.33%   | 707.812ms +0.50% | ±0.09% -59.38% |
| ParquetExtractorBench | bench_extract_10k | 1    | 3   | 239.705mb +0.01% | 1.008s -0.40%    | ±0.22% -51.98% |
| TextExtractorBench    | bench_extract_10k | 1    | 3   | 4.832mb +0.39%   | 23.314ms +1.54%  | ±1.09% +32.74% |
| XmlExtractorBench     | bench_extract_10k | 1    | 3   | 4.832mb +0.39%   | 558.392ms -0.52% | ±0.37% -88.07% |
+-----------------------+-------------------+------+-----+------------------+------------------+----------------+
Transformers 
+-----------------------------+--------------------------+------+-----+-----------------+-----------------+---------------+
| benchmark                   | subject                  | revs | its | mem_peak        | mode            | rstdev        |
+-----------------------------+--------------------------+------+-----+-----------------+-----------------+---------------+
| RenameEntryTransformerBench | bench_transform_10k_rows | 1    | 3   | 87.098mb +0.01% | 69.667ms -0.61% | ±0.63% +0.66% |
+-----------------------------+--------------------------+------+-----+-----------------+-----------------+---------------+
Loaders 
+--------------------+----------------+------+-----+------------------+------------------+----------------+
| benchmark          | subject        | revs | its | mem_peak         | mode             | rstdev         |
+--------------------+----------------+------+-----+------------------+------------------+----------------+
| AvroLoaderBench    | bench_load_10k | 1    | 3   | 95.281mb +0.02%  | 701.595ms -3.28% | ±0.82% -40.13% |
| CSVLoaderBench     | bench_load_10k | 1    | 3   | 46.114mb +0.04%  | 70.183ms +0.08%  | ±1.14% -7.07%  |
| JsonLoaderBench    | bench_load_10k | 1    | 3   | 90.559mb +0.02%  | 79.761ms -2.60%  | ±0.32% -80.61% |
| ParquetLoaderBench | bench_load_10k | 1    | 3   | 292.401mb +0.01% | 2.136s -1.79%    | ±0.28% -37.14% |
| TextLoaderBench    | bench_load_10k | 1    | 3   | 16.608mb +0.10%  | 36.980ms +0.98%  | ±0.71% -32.59% |
+--------------------+----------------+------+-----+------------------+------------------+----------------+
Building Blocks 
+-------------------------+----------------------------+------+-----+-----------------+------------------+-----------------+
| benchmark               | subject                    | revs | its | mem_peak        | mode             | rstdev          |
+-------------------------+----------------------------+------+-----+-----------------+------------------+-----------------+
| RowsBench               | bench_chunk_10_on_10k      | 2    | 3   | 60.708mb +0.00% | 4.010ms -2.59%   | ±2.25% -13.12%  |
| RowsBench               | bench_diff_left_1k_on_10k  | 2    | 3   | 80.500mb +0.00% | 175.837ms -0.72% | ±1.68% +224.43% |
| RowsBench               | bench_diff_right_1k_on_10k | 2    | 3   | 59.026mb +0.00% | 17.847ms -0.45%  | ±0.77% +10.90%  |
| RowsBench               | bench_drop_1k_on_10k       | 2    | 3   | 59.846mb +0.00% | 2.901ms -10.32%  | ±2.06% -29.09%  |
| RowsBench               | bench_drop_right_1k_on_10k | 2    | 3   | 59.846mb +0.00% | 2.861ms -14.11%  | ±0.78% -63.82%  |
| RowsBench               | bench_entries_on_10k       | 2    | 3   | 59.060mb +0.00% | 4.105ms -6.53%   | ±3.64% +100.23% |
| RowsBench               | bench_filter_on_10k        | 2    | 3   | 59.589mb +0.00% | 25.278ms -2.02%  | ±0.81% -49.76%  |
| RowsBench               | bench_find_on_10k          | 2    | 3   | 59.588mb +0.00% | 25.143ms -0.88%  | ±0.36% -65.07%  |
| RowsBench               | bench_find_one_on_10k      | 10   | 3   | 57.659mb +0.00% | 2.500μs -0.24%   | ±0.00% -100.00% |
| RowsBench               | bench_first_on_10k         | 10   | 3   | 57.659mb +0.00% | 0.500μs -16.67%  | ±0.00% 0.00%    |
| RowsBench               | bench_flat_map_on_1k       | 2    | 3   | 65.893mb +0.00% | 13.666ms -1.61%  | ±1.07% +51.99%  |
| RowsBench               | bench_map_on_10k           | 2    | 3   | 91.412mb +0.00% | 63.347ms -0.35%  | ±0.12% -82.85%  |
| RowsBench               | bench_merge_1k_on_10k      | 2    | 3   | 60.109mb +0.00% | 3.118ms -12.05%  | ±0.65% -53.35%  |
| RowsBench               | bench_partition_by_on_10k  | 2    | 3   | 62.379mb +0.00% | 49.746ms -2.41%  | ±0.44% -44.19%  |
| RowsBench               | bench_remove_on_10k        | 2    | 3   | 62.210mb +0.00% | 8.001ms -2.10%   | ±1.44% +9.50%   |
| RowsBench               | bench_sort_asc_on_1k       | 2    | 3   | 57.659mb +0.00% | 50.431ms -0.95%  | ±0.43% -24.36%  |
| RowsBench               | bench_sort_by_on_1k        | 2    | 3   | 57.659mb +0.00% | 49.379ms -1.80%  | ±0.48% +34.33%  |
| RowsBench               | bench_sort_desc_on_1k      | 2    | 3   | 57.659mb +0.00% | 51.578ms +0.76%  | ±1.57% +86.39%  |
| RowsBench               | bench_sort_entries_on_1k   | 2    | 3   | 59.934mb +0.00% | 9.364ms -0.62%   | ±0.50% +228.57% |
| RowsBench               | bench_sort_on_1k           | 2    | 3   | 57.659mb +0.00% | 37.685ms -0.36%  | ±0.25% -55.75%  |
| RowsBench               | bench_take_1k_on_10k       | 10   | 3   | 57.659mb +0.00% | 21.136μs -4.22%  | ±1.33% -29.13%  |
| RowsBench               | bench_take_right_1k_on_10k | 10   | 3   | 57.659mb +0.00% | 26.104μs -4.35%  | ±1.53% -22.58%  |
| RowsBench               | bench_unique_on_1k         | 2    | 3   | 80.500mb +0.00% | 179.595ms -1.13% | ±0.46% -53.77%  |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 93.715mb +0.00% | 155.361ms -0.24% | ±0.44% +16.19%  |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 48.630mb +0.01% | 79.077ms -1.75%  | ±0.19% -60.57%  |
| NativeEntryFactoryBench | bench_entry_factory        | 1    | 3   | 12.653mb -0.02% | 18.607ms -3.64%  | ±0.31% -67.82%  |
+-------------------------+----------------------------+------+-----+-----------------+------------------+-----------------+

@stloyd
Copy link
Member

stloyd commented Nov 12, 2023

@norberttech I think that it's leftover as the implementation started to be similar to ElasticSearch, but MeiliSearch has "better" handling of updates for multiple documents as it uses the same endpoint as for new ones.

I would leave it and open a new issue to validate if we want to have:

  • add new and update existing with given IDs (current),
  • or add new and replace existing with given IDs,

@norberttech norberttech merged commit 2832e03 into flow-php:1.x Nov 12, 2023
@norberttech norberttech deleted the 738-apply-default-batchsize-based-on-loader-type branch December 5, 2024 20:24
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

1 more reviewer

@stloyd stloyd stloyd approved these changes

Reviewers whose approvals may not affect merge requirements

Assignees

No one assigned

Labels

size: L

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

Apply default batchSize based on Loader type

2 participants

@norberttech @stloyd