Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/freymaurer/pysafe
A web service for ml prediction. Using SAFE stack, fastapi, websockets, daisyui.
https://github.com/freymaurer/pysafe
aspnetcore daisyui data-science fable fastapi fsharp ml python safe-stack
Last synced: about 1 month ago
JSON representation
A web service for ml prediction. Using SAFE stack, fastapi, websockets, daisyui.
- Host: GitHub
- URL: https://github.com/freymaurer/pysafe
- Owner: Freymaurer
- Created: 2024-04-16T06:49:36.000Z (8 months ago)
- Default Branch: main
- Last Pushed: 2024-05-13T14:42:34.000Z (7 months ago)
- Last Synced: 2024-11-21T04:37:03.280Z (about 1 month ago)
- Topics: aspnetcore, daisyui, data-science, fable, fastapi, fsharp, ml, python, safe-stack
- Language: F#
- Homepage:
- Size: 67.4 MB
- Stars: 0
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# PySAFE
# Local Development
## Install pre-requisites
You'll need to install the following pre-requisites in order to build SAFE applications
* [.NET SDK](https://www.microsoft.com/net/download) 8.0 or higher
* [Node 18](https://nodejs.org/en/download/) or higher
* [NPM 9](https://www.npmjs.com/package/npm) or higher
* [Python 3.10](https://www.python.org/downloads/) or higher
## Install
- run `setup.cmd`
.. __or__ ..
1. `dotnet tool restore`
2. `py -m venv .venv`
3. `.\.venv\Scripts\python.exe -m pip install -r .\src\FastAPI\requirements.txt`
## Run
- `.\build.cmd run` starts SAFE stack
plus in another terminal run:
1. activate local python environment: `.\.venv\Scripts\Activate.ps1`
2. navigate to fastapi folder: `cd .\src\FastAPI\`
3. start fastapi backend: `python -m uvicorn app.main:app --reload`
## Activate Email notification (optional)
Set user-secrets in the following schema:
```json
{
"email": {
"NET_EMAIL_EMAIL": "[email protected]",
"NET_EMAIL_ACCOUNTNAME": "PlaceholderAccountName",
"NET_EMAIL_PASSWORD": "HelloWorld1234",
"NET_EMAIL_SERVER": "smtp.placeholdermail.de",
"NET_EMAIL_PORT": 587
}
}
```
# Request Workflow
```mermaid
sequenceDiagram
participant py as Python ML
participant net as F#35; Server
participant c as Client
actor u as User
u -->> c: Gives data
c -->>+net: sends user data
par start analysis
net-)+py: sends data, trigger eval
py-)net: returns binned data
and return request information
net -) c: returns `request-ID`
end
critical ⚠️
u -->> c: copies and stores `request-ID`
end
opt email
u -->> c: give email address
c -->> net: give id + email to store
end
opt check status
u -->> c: use `request-ID` to check status
end
py-)net: send last package
deactivate py
net-->>net: run q-value calculation
net-->>net: store results
deactivate net
opt gave email
net-)u: send email
end
u -->> c: request data
c-->>net: get data
net-->>c: return data
c-->>u: download data
```
# Result
Explanations of Chloropred ,Qchloro, Mitopred,Qmito,Secrpred,Qsecr, and FinalPred.
### Chloropred
Prediction score indicating the likelihood of the protein being localized to the Chloroplast. A higher scores suggest a stronger prediction that the protein is localized in the Chloroplast.
### Qchloro
q-value associated with the Chloroplast prediction score. Provides a measure of statistical significance for the Chloroplast prediction. Lower q-values indicate higher statistical significance.
### Mitopred
Prediction score for the localization of the protein to the Mitochondria. A higher scores suggest a stronger prediction of Mitochondrial localization.
### Qmito
q-value associated with the Mitochondria prediction score. Indicates the statistical significance of the Mitochondria localization prediction. Lower q-values suggest a more reliable prediction.
### Secrpred
Prediction score for identifying the protein as a Secretory Protein.A higher scores indicate a stronger likelihood that the protein functions as a Secretory Protein.
### Qsecr
q-value for the Secretory Protein prediction. Provides a measure of the statistical significance of the Secretory Protein prediction. Lower q-values are indicative of more statistically significant predictions.
### FinalPred
Represents the model's final prediction of the protein's localization based on the highest score and its corresponding q-value. The final localization is determined by comparing the q-values and prediction scores against preset cutoffs. If all q-values exceed the cutoff, the protein is classified as "Cytoplasmic."
### Cutoff
The threshold q-value below which a prediction is considered statistically significant. Set to 0.05 by default, meaning that predictions with q-values below this threshold are classified as significant. This parameter helps in distinguishing between statistically significant and non-significant predictions, reducing the chance of false-positive localizations.