No project description provided
Project description
algovault
Experimentation tracking
Features
- Fully embrace the relational model
- i.e. run is many-to-many with experiment, so reuse of run results across experiments is possible
- all operations idempotent
- Suitable for use in a workflow orchestration system
- built-in checking for presence of results
- can be used to cache runs
- high performance read and write
- I'm looking at you, mlflow.get_metric_history
- dead-simple integration points
- The data model is simply sqlite files
- serverless
- built-in aggregations
- computing common aggregates is crazy fast and requires little memory
- no magic
- no global context means you can paralellize fearlessly
Design
- writers write to a local copy of sqlite database (maybe in-memory?)
- runs end and the databases sent to checkpoint location
- upon read, compact checkpoints to a single database instance (read replica)
- read replica knows which instances have already been ingested, incremental update
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
algovault-0.0.3.tar.gz
(17.2 kB
view hashes)
Built Distributions
Close
Hashes for algovault-0.0.3-pp39-pypy39_pp73-manylinux_2_5_x86_64.manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 8f41f824a9860c8c544bd90f91394f614f72b7ceab785769fd480e7215cef0d4 |
|
MD5 | 12e95a69563b9a1b02d08616e838ebf2 |
|
BLAKE2b-256 | ccb390e1c74ceff8a225d155d131b9042461ff5a72df6af69295daf9e93f7718 |
Close
Hashes for algovault-0.0.3-pp38-pypy38_pp73-manylinux_2_5_x86_64.manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 745d96b21fd6446c24385e895c7c10914bcaad882105380bde6b727f4fee9af7 |
|
MD5 | 68ba859633838187984b3cb89dd076b6 |
|
BLAKE2b-256 | e63492dcacdae1d9311ad79f9e648da535bc4acb6647b3c4853d536ddfc76c10 |
Close
Hashes for algovault-0.0.3-pp37-pypy37_pp73-manylinux_2_5_x86_64.manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 3e88b2274b0c7ff828edbdc0587767ab2fd2f545f488daeaae0422f29263a54c |
|
MD5 | 1433f0ba741196250772108c012de685 |
|
BLAKE2b-256 | 928e4c48a82146b5afaf1de0e29565c18187162d6510a9c4ef76d8db59810de2 |
Close
Hashes for algovault-0.0.3-cp311-cp311-manylinux_2_5_x86_64.manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 0e2beeaa5f02a84b9c6d0cb2949a69a6a8c50e6cc652b3d3a1e7508341e65dc2 |
|
MD5 | 79addb4128a71774bf745f38361f82d1 |
|
BLAKE2b-256 | 33038663707321c7eeff9061585b6c3f5b8678f169ee3e30b9d6c93f5fe038c0 |
Close
Hashes for algovault-0.0.3-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 7c77b8d7fa0b66bba2c03f84102a17e09111bcedf96f0477205ffce681995fb1 |
|
MD5 | 1dd189b979eb92211ac2c8281057cb83 |
|
BLAKE2b-256 | 2e9327ced5aef3e126be34d3c33ccfbdb3e90be700ef251cc59a40ddfbd20eec |
Close
Hashes for algovault-0.0.3-cp39-cp39-manylinux_2_5_x86_64.manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | e213a7562f9fa3c73d25eeda5b66bcf297a9e2929076e487c8042da800ed2c00 |
|
MD5 | b1f84a5edd55f3f725465adbfb7b6073 |
|
BLAKE2b-256 | 883c032e7b7fef49c81f2cb270769d0fea5168dfd0a2fdb3afedd9f78642e042 |
Close
Hashes for algovault-0.0.3-cp38-cp38-manylinux_2_5_x86_64.manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 2de3a32e539fc976894c43e49d02c702bfebe6022143cf36f3349430e3636686 |
|
MD5 | 982e5a46638a1eaf467dc2ee299ac259 |
|
BLAKE2b-256 | 47bcee4327e3773503bdc4917b12d39b4fd0d89f926c6762e23ddb0936bc7ea4 |
Close
Hashes for algovault-0.0.3-cp37-cp37m-manylinux_2_5_x86_64.manylinux1_x86_64.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | 9c73bd36693ff353a1d333e59cfbafac2023d289cd7e491a26af7eee91b33320 |
|
MD5 | bf90f69b047c075bce318ee75d88c072 |
|
BLAKE2b-256 | 88a33f699969f5b6b7aeaee9a06359221fbaf6765cd41be441e5e70d0fffb87d |