Skip to content

Commit 7cafc2f

Browse files
authored
17455 revamp extras (#7)
This PR: - Support for additional `extras` installed into `howso/resources/` to combined multiple `extras.yml` into one, merged set of `client_extra_config` passed into the client initializer. - Updates the README.md to include a better SVG image that should work better with PyPI.org - Updates the description in the pyproject.toml file to better reflect Howso's vision
1 parent 30a1ab6 commit 7cafc2f

File tree

8 files changed

+145
-151
lines changed

8 files changed

+145
-151
lines changed

.gitignore

+1-1
Original file line numberDiff line numberDiff line change
@@ -8,11 +8,11 @@
88
**/*.trace
99
dumps
1010
traces
11-
howso/direct/resources
1211
howso/docs/build
1312
howso/docs/source/_build
1413
howso/docs/source/auto_examples
1514
howso/howso-engine
15+
howso/resources
1616
local/
1717

1818
# Do not ignore the thumbnails used by sphinx-gallery

HowsoLogo.svg

-16
This file was deleted.

MANIFEST.in

-1
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,6 @@
11
include README.md
22
global-include requirements-*.txt
33
global-include requirements.in
4-
global-include HowsoLogo.svg
54
include LICENSE.txt
65
include LICENSE-3RD-PARTY.txt
76
include version.json

README.md

+53-21
Original file line numberDiff line numberDiff line change
@@ -1,29 +1,55 @@
11

22
<div align="left">
3-
<img src="HowsoLogo.svg">
3+
<img src="https://howso.com/assets/images/Howso_Readme.svg">
44
</div>
55

6-
The Howso Engine&trade; is a natively and fully explainable ML engine, serving as an alternative to black box AI neural networks. It’s core features give users data exploration and machine learning capabilities through the creation and use of Trainees that help users store, explore, and analyze the relationships in their data. Howso&trade; leverages an instance-based learning approach with strong ties to the [k-nearest neighbors algorithm](https://en.wikipedia.org/wiki/K-nearest_neighbors_algorithm) and [information theory](https://en.wikipedia.org/wiki/Information_theory) to scale for real world applications.
7-
8-
At the core of Howso is the concept of a Trainee, a collection of cases that comprise knowledge. In traditional ML, this is typically referred to as a model, but a Trainee may additionally include metadata, parameters, details of feature attributes, with data lineage and provenance. Unlike traditional ML, Trainees are designed to be versatile, a single model that after training a dataset can do the following without the need to retrain:
9-
- Perform **classification** on any target feature using any set of input features
6+
The Howso Engine&trade; is a natively and fully explainable ML engine, serving
7+
as an alternative to black box AI neural networks. Its core features give users
8+
data exploration and machine learning capabilities through the creation and use
9+
of Trainees that help users store, explore, and analyze the relationships in
10+
their data. Howso&trade; leverages an instance-based learning approach with
11+
strong ties to the [k-nearest neighbors
12+
algorithm](https://en.wikipedia.org/wiki/K-nearest_neighbors_algorithm) and
13+
[information theory](https://en.wikipedia.org/wiki/Information_theory) to scale
14+
for real world applications.
15+
16+
At the core of Howso is the concept of a Trainee, a collection of cases that
17+
comprise knowledge. In traditional ML, this is typically referred to as a model,
18+
but a Trainee may additionally include metadata, parameters, details of feature
19+
attributes, with data lineage and provenance. Unlike traditional ML, Trainees
20+
are designed to be versatile, a single model that after training a dataset can
21+
do the following without the need to retrain:
22+
- Perform **classification** on any target feature using any set of
23+
input features
1024
- Perform **regression** on any target feature using any set of input features
1125
- Perform **anomaly detection** based on any set of features
1226
- Measure **feature importance** for predicting any target feature
13-
- **Synthesize** data that maintains the same feature relationships of the original data while maintaining privacy
27+
- **Synthesize** data that maintains the same feature relationships of the
28+
original data while maintaining privacy
1429

1530
Furthermore, Trainees are auditable, debuggable, and editable.
16-
- **Debuggable**: Every prediction of a Trainee can be drilled down to investigate which cases from the training data were used to make the prediction.
17-
- **Auditable**: Trainees manage metadata about themselves including: when data is trained, when training data is edited, when data is removed, etc.
18-
- **Editable**: Specific cases of training data can be removed, edited, and emphasized (through case weighting) without the need to retrain.
31+
- **Debuggable**: Every prediction of a Trainee can be drilled down to
32+
investigate which cases from the training data were used to make
33+
the prediction.
34+
- **Auditable**: Trainees manage metadata about themselves including: when data
35+
is trained, when training data is edited, when data is removed, etc.
36+
- **Editable**: Specific cases of training data can be removed, edited, and
37+
emphasized (through case weighting) without the need to retrain.
1938

2039
## Resources
2140
- [Documentation](https://docs.howso.com)
2241
- [Howso Engine Recipes (sample notebooks)](https://github.com/howsoai/howso-engine-recipes)
2342
- [Howso Playground](https://playground.howso.com)
2443

2544
## General Overview
26-
This Repo provides the Python interface with [Howso Engine](https://github.com/howsoai/howso-engine) that exposes the Howso Engine functionality. The Client objects directly interface with the engine API endpoints while the Trainee objects provides the python functionality for general users. Client functions may be called by the user but for most workflows the Trainee functionality is sufficient. Each Trainee represents an individual Machine Learning object or model that can perform functions like training and predicting, while a client may manage the API interface for multiple Trainees.
45+
This Repo provides the Python interface with
46+
[Howso Engine](https://github.com/howsoai/howso-engine) that exposes the Howso
47+
Engine functionality. The Client objects directly interface with the engine API
48+
endpoints while the Trainee objects provides the python functionality for
49+
general users. Client functions may be called by the user but for most workflows
50+
the Trainee functionality is sufficient. Each Trainee represents an individual
51+
Machine Learning object or model that can perform functions like training and
52+
predicting, while a client may manage the API interface for multiple Trainees.
2753

2854

2955
## Supported Platforms
@@ -46,16 +72,20 @@ To install the current release:
4672
pip install howso-engine
4773
```
4874

49-
You can verify your installation is working by running the following command in your python environment terminal:
75+
You can verify your installation is working by running the following command in
76+
your python environment terminal:
5077

5178
```bash
5279
verify_howso_install
5380
```
5481

55-
See the Howso Engine [Install Guide](https://docs.howso.com/getting_started/installing.html) for additional help and troubleshooting information.
82+
See the Howso Engine
83+
[Install Guide](https://docs.howso.com/getting_started/installing.html) for
84+
additional help and troubleshooting information.
5685
## Usage
5786

58-
The Howso Engine is designed to support users in the pursuit of many different machine learning tasks using Python.
87+
The Howso Engine is designed to support users in the pursuit of many different
88+
machine learning tasks using Python.
5989

6090
Below is a very high-level set of steps recommended for using the Howso Engine:
6191

@@ -65,11 +95,13 @@ Below is a very high-level set of steps recommended for using the Howso Engine:
6595
4. Call Analyze on the Trainee to find optimal hyperparameters
6696
5. Explore your data!
6797

68-
Once the Trainee has been given feature attributes, trained, and analyzed, then the Trainee is ready
69-
to be used for all supported machine learning tasks. At this point one could start making predictions
70-
on unseen data, investigate the most noisy features, find the most anomalous training cases, and much more.
98+
Once the Trainee has been given feature attributes, trained, and analyzed, then
99+
the Trainee is ready to be used for all supported machine learning tasks. At
100+
this point one could start making predictions on unseen data, investigate the
101+
most noisy features, find the most anomalous training cases, and much more.
71102

72-
Please see the [User Guide](https://docs.howso.com/user_guide/index.html) for basic workflows as well as additional information about:
103+
Please see the [User Guide](https://docs.howso.com/user_guide/index.html) for
104+
basic workflows as well as additional information about:
73105
- Anomaly detection
74106
- Classification
75107
- Regression
@@ -83,8 +115,10 @@ Please see the [User Guide](https://docs.howso.com/user_guide/index.html) for ba
83115
- Trainee editing
84116
- ID-based privacy
85117

86-
There is also a set of basic [Python examples](https://docs.howso.com/examples/index.html) that provide a complete set of Python scripts and [Jupyter notebooks](https://jupyter.org/) to run.
87-
118+
There is also a set of basic [Jupyter notebooks](https://jupyter.org/) to run
119+
that provides a
120+
[complete set of examples](https://docs.howso.com/examples/index.html) of how
121+
to use Howso Engine.
88122

89123
## License
90124

@@ -93,5 +127,3 @@ There is also a set of basic [Python examples](https://docs.howso.com/examples/i
93127
## Contributing
94128

95129
[Contributing](CONTRIBUTING.md)
96-
97-

0 commit comments

Comments
 (0)