### TPU/CPU/GPU Strategy Setup (Python)

Source: https://documentation.picsellia.com/docs/final-script

Configures the distributed training strategy, attempting to connect to a TPU runtime if available, otherwise falling back to a CPU/GPU strategy.

```python
##Harware
try:
    tpu = tf.distribute.cluster_resolver.TPUClusterResolver.connect()
    print("Device:", tpu.master())
    strategy = tf.distribute.TPUStrategy(tpu)
except ValueError:
    print("Not connected to a TPU runtime. Using CPU/GPU strategy")
    strategy = tf.distribute.MirroredStrategy()
```

--------------------------------

### Image Preprocessing and Augmentation Setup (Python)

Source: https://documentation.picsellia.com/docs/final-script

Sets up image data generators for training and testing, including rescaling and augmentation. It also defines an image augmentation pipeline using Keras layers.

```python
##Preprocessing
size = (IMG_SIZE, IMG_SIZE)

train_img_gen = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1./255, rotation_range=20)
test_img_gen = tf.keras.preprocessing.image.ImageDataGenerator(rescale=1./255)
ds_train = next(train_img_gen.flow_from_directory('train', target_size=size, batch_size=batch_size, seed=42))
ds_test = next(test_img_gen.flow_from_directory('test', target_size=size, batch_size=batch_size, seed=42))


NUM_CLASSES = len(experiment.get_dataset('train').list_labels())

img_augmentation = Sequential(
    [
        layers.RandomRotation(factor=0.15),
        layers.RandomTranslation(height_factor=0.1, width_factor=0.1),
        layers.RandomFlip(),
        layers.RandomContrast(factor=0.1),
    ],
    name="img_augmentation",
)

print("Dataloader Instantiated")
```

--------------------------------

### Setup Feedback Loop for Picsellia Deployment

Source: https://documentation.picsellia.com/reference/deployment

Sets up the Feedback Loop for a Picsellia Deployment. A Dataset Version can be specified during setup or attached later using attach_dataset_to_feedback_loop(). This feature helps in augmenting the training set with quality data.

```python
deployment.setup_feedback_loop(dataset_version)
```

--------------------------------

### Setup Continuous Deployment

Source: https://documentation.picsellia.com/reference/deployment

Configures the continuous deployment policy for a deployment.

```APIDOC
## POST /websites/picsellia/setup_continuous_deployment

### Description
Set up the continuous deployment for this pipeline.

### Method
POST

### Endpoint
/websites/picsellia/setup_continuous_deployment

### Parameters
#### Request Body
- **policy** (string or ContinuousDeploymentPolicy enum) - Required - The policy to use for continuous deployment (e.g., 'DEPLOY_MANUAL').

### Request Example
```json
{
  "policy": "DEPLOY_MANUAL"
}
```

### Response
#### Success Response (200)
- **message** (string) - Confirmation message indicating successful setup.

#### Response Example
```json
{
  "message": "Continuous deployment setup complete."
}
```
```

--------------------------------

### Dockerfile Setup for Picsellia

Source: https://documentation.picsellia.com/docs/part-4-make-your-modelversion-trainable

This Dockerfile demonstrates how to set up a custom Docker image for Picsellia. It specifies a base CUDA image, installs Python dependencies from a requirements file, and sets the working directory and entrypoint for the training script. Ensure the CUDA version is compatible with your hardware, especially Tesla V100s GPUs.

```dockerfile
FROM picsellia/cuda:11.7.1-cudnn8-ubuntu20.04

COPY your-custom-image/requirements.txt .
RUN pip install -r requirements. txt

# You can also put the picsellia package inside your requirements
RUN pip install -r picsellia --upgrade

WORKDIR /picsellia

COPY your-custom-image/picsellia .

ENTRYPOINT ["run", "main.py"]
```

--------------------------------

### Install Picsellia Python Package

Source: https://documentation.picsellia.com/docs/annotation-formats

Installs the Picsellia Python package, which is required for importing annotations into the Picsellia platform. This is a prerequisite for using the provided Python code examples.

```bash
pip install picsellia
```

--------------------------------

### Python Training Script Initialization with Picsellia Client

Source: https://documentation.picsellia.com/docs/part-4-make-your-modelversion-trainable

This Python script snippet shows the initial setup for integrating Picsellia into a training script. It imports necessary libraries, initializes the Picsellia Client with API token, organization name, and host, and retrieves a specific project and experiment. This setup is a prerequisite for sending training logs and monitoring progress via the Picsellia platform.

```python
from tensorflow.keras.applications import EfficientNetB0
from tensorflow.keras.models import Sequential
from tensorflow.keras import layers
from tensorflow import keras
import tensorflow as tf
import tensorflow_datasets as tfds
from picsellia import Client, Experiment
from picsellia.types.enums import LogType
import os
import numpy as np 

print("Initializing Picsellia Client 🥑")

client = Client(
  api_token="XXXXXXX", 
  organization_name="my_organization",
  host="https://app.picsellia.com"
)
project = client.get_project("documentation_project")
experiment = project.get_experiment("exp-0-documentation")
datasets = experiment.list_attached_dataset_versions()


# The rest of your script
```

--------------------------------

### Job - start_logging_buffer

Source: https://documentation.picsellia.com/reference/job

Starts a logging buffer to accumulate log entries before sending them.

```APIDOC
## Job - start_logging_buffer

### Description
Start a logging buffer.

### Method
POST

### Endpoint
/job/{job_id}/log/buffer/start

### Parameters
#### Request Body
- **length** (int) - Optional - Buffer length. Defaults to 1.

### Request Example
```json
{
  "length": 10
}
```

### Response
#### Success Response (200)
- **message** (string) - Confirmation message.

#### Response Example
```json
{
  "message": "Logging buffer started successfully."
}
```
```

--------------------------------

### Job - start_logging_chapter

Source: https://documentation.picsellia.com/reference/job

Starts a new logging chapter, allowing for organized log entries.

```APIDOC
## Job - start_logging_chapter

### Description
Start a logging chapter.

### Method
POST

### Endpoint
/job/{job_id}/log/chapter/start

### Parameters
#### Request Body
- **name** (str) - Required - Chapter name.

### Request Example
```json
{
  "name": "chapter1"
}
```

### Response
#### Success Response (200)
- **message** (string) - Confirmation message.

#### Response Example
```json
{
  "message": "Logging chapter started successfully."
}
```
```

--------------------------------

### Picsellia Client Initialization and Data Retrieval (Python)

Source: https://documentation.picsellia.com/docs/final-script

Initializes the Picsellia client, retrieves a project and experiment, and downloads attached dataset versions and their associated assets. It also downloads model weights and training parameters.

```python
from tensorflow.keras.applications import EfficientNetB0
from tensorflow.keras.models import Sequential
from tensorflow.keras import layers
from tensorflow import keras
import tensorflow as tf
import tensorflow_datasets as tfds
from picsellia import Client, Experiment
from picsellia.types.enums import LogType
import os
import numpy as np 

##Client initialization
print("Initializing Picsellia Client 🥑")

client = Client(
  api_token="XXXXXXX", 
  organization_name="my_organization",
  host="https://app.picsellia.com"
)

##Retrieve experiment to train
project = client.get_project("documentation_project")
experiment = project.get_experiment("exp-0-documentation")


##Retreive & download datasets
datasets = experiment.list_attached_dataset_versions()
print("Downloading Images for datasets ...")

for dataset in datasets:
    for label in dataset.list_labels():
        os.makedirs(os.path.join(dataset.version, label.name))
        assets = dataset.list_assets(q=f"annotations.classifications.label.name = \"{str(label.name)}\"").download(os.path.join(dataset.name, label.name))

print("Downloading Images for datasets ... ✅")


##Retrieve and download model weights
print("Downloading Weights for Model ... ")

base_model = experiment.get_base_model_version()
base_model_weights = base_model.get_file('weights')
base_model_weights.download()

print("Downloading Weights for Model ... ✅")

##Retrieve experiment training parameters
parameters = experiment.get_log(name='parameters').data
IMG_SIZE = parameters['image_size']
batch_size = parameters['batch_size']
```

--------------------------------

### POST /setup_feedback_loop

Source: https://documentation.picsellia.com/reference/deployment

Set up the Feedback Loop for a Deployment. You can optionally attach a Dataset Version during setup or use `attach_dataset_version_to_feedback_loop` afterward. This is useful for augmenting your training set with quality data.

```APIDOC
## POST /setup_feedback_loop

### Description
Set up the Feedback Loop for a Deployment. You can specify one Dataset Version to attach to it or use the attach_dataset_to_feedback_loop() afterward, so you can add multiple ones. This is a great option to increase your training set with quality data.

### Method
POST

### Endpoint
/setup_feedback_loop

### Parameters
#### Path Parameters
None

#### Query Parameters
None

#### Request Body
- **dataset_version** (DatasetVersion | None) - Optional - The Dataset Version to attach. Note: This parameter is deprecated. Use `attach_dataset_version_to_feedback_loop()` instead.

### Request Example
```python
dataset_version = client.get_dataset("my-dataset").get_version("latest")
deployment = client.get_deployment(
    name="awesome-deploy"
)
deployment.setup_feedback_loop(dataset_version)
```

### Response
#### Success Response (200)
- **message** (str) - Confirmation message indicating the feedback loop setup.

#### Response Example
```json
{
  "message": "Feedback loop setup successfully."
}
```
```

--------------------------------

### HuggingFace Zero-Shot Classifier Example

Source: https://documentation.picsellia.com/docs/evaluate-a-classification-model

This Python code demonstrates how to use a Zero-Shot classifier (CLIP model) from HuggingFace. It loads a pre-trained model and processor, fetches an image from a URL, and then uses the processor to get similarity scores between the image and provided text labels.

```python
from PIL import Image
import requests

from transformers import CLIPProcessor, CLIPModel

model = CLIPModel.from_pretrained("openai/clip-vit-large-patch14")
processor = CLIPProcessor.from_pretrained("openai/clip-vit-large-patch14")

url = "http://images.cocodataset.org/val2017/000000039769.jpg"
image = Image.open(requests.get(url, stream=True).raw)

inputs = processor(text=["a photo of a cat", "a photo of a dog"], images=image, return_tensors="pt", padding=True)

outputs = model(**inputs)
logits_per_image = outputs.logits_per_image # this is the image-text similarity score
probs = logits_per_image.softmax(dim=1) # we can take the softmax to get the label probabilities
```

--------------------------------

### Setup Continuous Training Pipeline (Python)

Source: https://documentation.picsellia.com/reference/deployment

Initializes and activates continuous training features. A training will be triggered based on the attached dataset versions when the deployment pipeline hits a specified trigger. Supports experiment parameters but not scan configurations at the moment. Requires a Project object and optionally accepts dataset_version, model_version, trigger, threshold, and experiment_parameters.

```python
deployment.setup_continuous_training(
    project, threshold=150, experiment_parameters=experiment_parameters
)
```

--------------------------------

### Create Campaign for Deployment (Python)

Source: https://documentation.picsellia.com/reference/deployment

Creates a new campaign for a deployment. Allows setting a description, instructions (from file or text), end date, and flags for automatically adding new assets or closing the campaign upon completion. Returns a ReviewCampaign object.

```python
def create_campaign(
   description: (str|None) = None, instructions_file_path: (str|None) = None,
   instructions_text: (str|None) = None, end_date: (date|None) = None,
   auto_add_new_assets: (bool|None) = False,
   auto_close_on_completion: (bool|None) = False
)
```

Create campaign on a deployment.

```python
foo_deployment.create_campaign()
```
```

--------------------------------

### Call Picsellia Monitoring Service with Python

Source: https://documentation.picsellia.com/docs/predicting-and-monitoring-without-python-sdk

This Python snippet demonstrates how to connect to the Picsellia monitoring service. It includes a function to generate a JWToken for authentication and an example of sending prediction data via an HTTP POST request. Ensure you have the 'requests' library installed. The function requires the monitoring service host, deployment ID, and your API token. It outputs prediction results from the monitoring service.

```python
import base64
import mimetypes

import requests
import json


def generate_authorization_header(host: str, deployment_id: str, api_token: str):
    """This method calls our monitoring service to create a jwt and build authorization header.
    Our jwt are usually available for 1 hour, you might need to recall this method at some point to maintain connection.

    :param host: of our monitoring service
    :param deployment_id: deployment identifier that you want to use
    :param api_token: your picsellia api token
    :return: a dict representing headers with a JWT of connection to our service.
    """
    # Take care, this auth url is different from prediction service one
    auth_url = host + "/api/auth/login"
    auth_headers = {
        "Authorization": "Token " + api_token,
    }
    # json.dumps() serialize a python dictionary to a json string
    auth_payload = json.dumps({"deployment_id": deployment_id, "api_token": api_token})

    # requests.post() makes an HTTP 'POST', on given url, with given headers, and given json payload
    # calling .json() return the response as a dict, that have the wanted key "jwt"
    response = requests.post(
        url=auth_url, headers=auth_headers, data=auth_payload
    ).json()
    return {
        "Authorization": "Bearer " + response["jwt"],
    }


if __name__ == "__main__":
    # Host of our picsellia prediction service.
    host = "https://serving.picsellia.com"

    # Fill these variables with your data
    deployment_id = "<YOUR_DEPLOYMENT_ID>"
    api_token = "<YOUR_PICSELLIA_API_TOKEN>"

    # Generate authorization headers by calling our prediction service
    headers = generate_authorization_header(host, deployment_id, api_token)

    # Url of our service to make a prediction
    url = host + "/api/deployment/" + deployment_id + "/predictions"

    # This block is needed in our snippet, because we want to find content type, filename
    # also we have to encode image to a base64 data : monitoring api only allows this type of encoding at the moment
    # image_width and image_height are not computed in this snippet but need to be sent to monitoring service
    # You might already have opened this image, you might want to do it differently
    image_path = "/path/to/your/image"
    image_height = 600
    image_width = 800
    with open(image_path, "rb") as img_file:
        content_type = mimetypes.guess_type(image_path, strict=False)[0]
        filename = image_path.split("/")[-1]
        encoded_image = base64.b64encode(img_file.read()).decode("utf-8")

    # latency given by your model to predict on this image
    latency = 0.259

    # prediction is a python dictionary representing shapes predicted by your model,
    # it needs different keys depending on your model. For example for an ObjectDetection model:
    raw_predictions = {
        "detection_scores": [
            0.7279638051986694,  # first shape's score and should be between 0 and 1
            0.7203983664512634,
            0.7142382264137268,
        ],
        "detection_boxes": [
            [57, 213, 118, 256],  # first shape's box as [y1, x1, y2, x2] coordinates
            [186, 165, 246, 248],
            [80, 73, 169, 199],
        ],
        "detection_classes": [
            3,  # first shape's class, please make sure your labelmap is well-defined
            3,
            4,
        ],
    }

    data = json.dumps(
        {
            "filename": filename,
            "content_type": content_type,
            "height": image_height,
            "width": image_width,
            "image": encoded_image,
            "raw_predictions": raw_predictions,
            "latency": latency,
        }
    )
    # requests.post() will send your prediction to our monitoring service
    response = requests.post(url=url, headers=headers, data=data)

    monitoring_response = response.json()
    print(monitoring_response)

```

--------------------------------

### Start Logging Buffer (Python)

Source: https://documentation.picsellia.com/reference/experiment

Starts a logging buffer for an experiment. This can be used to group log entries. The buffer length can be specified, defaulting to 1.

```python
experiment.start_logging_buffer()
```

--------------------------------

### Initialize Picsellia Client in Trial Environment (Python)

Source: https://documentation.picsellia.com/docs/installation

This Python code snippet demonstrates how to initialize the Picsellia client for the trial environment. It requires your organization name, API token, and the trial host URL.

```python
# Organization name
organization_name = "organization_name"

# Define the environement to connect to
host = "https://trial.picsellia.com"

# Client initialization in the Trial environment
client = Client(api_token="YOUR API TOKEN", organization_name=organization_name, host=host)
```

--------------------------------

### Install MLFlow and Picsellia Dependencies

Source: https://documentation.picsellia.com/docs/use-mlflow-with-picsellia

Installs necessary Python libraries for MLFlow, Picsellia, PyTorch, and related utilities using pip. This command ensures all required packages are available for the experiment.

```python
%pip install -q mlflow databricks-sdk torchmetrics torchinfo picsellia

```

--------------------------------

### Create Deployment from Model Version

Source: https://documentation.picsellia.com/reference/client

Creates a deployment from a given model version. This allows for monitoring and potentially a hosted endpoint. It accepts optional parameters for shadow model versions, deployment name, detection thresholds, target datalake, and serving disablement.

```python
model_version = client.get_model_version_by_id('918351d2-3e96-4970-bb3b-420f33ded895')
deployment = client.create_deployment(model_version=model_version)
```

--------------------------------

### Get Data Tag

Source: https://documentation.picsellia.com/reference/datalake

Retrieves a data tag by its name.

```APIDOC
## GET /websites/picsellia/data/tags/{name}

### Description
Retrieves a data tag by its name.

### Method
GET

### Endpoint
/websites/picsellia/data/tags/{name}

### Parameters
#### Path Parameters
- **name** (str) - Required - Name of the tag to retrieve.

### Response
#### Success Response (200)
- **Tag** (object) - A Tag object representing the retrieved tag.

#### Response Example
{
  "example": "Tag object"
}
```

--------------------------------

### Full Training Script Initialization with Picsellia Data

Source: https://documentation.picsellia.com/docs/part-3-retrieve-data-files-and-parameters-from-picsellia

This comprehensive Python script integrates Picsellia SDK functionalities to initialize a client, retrieve project and experiment details, download datasets, fetch model weights, and load training parameters. It sets up the necessary data and configurations for a machine learning training process.

```python
from tensorflow.keras.applications import EfficientNetB0
from tensorflow.keras.models import Sequential
from tensorflow.keras import layers
import tensorflow as tf
import tensorflow_datasets as tfds
from picsellia import Client
import os

## Initialization of the Picsellia client
client = Client(
  api_token="XXXXXXX", 
  organization_name="my_organization",
  host="https://app.picsellia.com"
)

## Retrieve the Picsellia experiment that needs to be trained
project = client.get_project("documentation_project")
experiment = project.get_experiment("exp-0-documentation")

## Download the datasets from Picsellia and organize the data on the infrastrcture drive
datasets = experiment.list_attached_dataset_versions()
for dataset in datasets:
    for label in dataset.list_labels():
        os.makedirs(os.path.join(dataset.version, label.name))
        assets = dataset.list_assets(q=f"annotations.classifications.label.name = \"{str(label.name)}\"").download(os.path.join(dataset.name, label.name))

## Get base model of this experiment and download associated files
base_model = experiment.get_base_model_version()
base_model_weights = base_model.get_file('weights')
base_model_weights.download()

## Get experiment parameters that will define the training
parameters = experiment.get_log(name='parameters').data
IMG_SIZE = parameters['image_size']
batch_size = parameters['batch_size']

```

--------------------------------

### Get Processing API

Source: https://documentation.picsellia.com/reference/client

Retrieves a specific processing by its name.

```APIDOC
## GET /websites/picsellia/processings/{name}

### Description
Find a processing in this organization by its unique name.

### Method
GET

### Endpoint
/websites/picsellia/processings/{name}

### Parameters
#### Path Parameters
- **name** (str) - Required - The name of the processing to retrieve.

### Request Example
```
GET /websites/picsellia/processings/auto-tagging-dataset
```

### Response
#### Success Response (200)
- **processing** (Processing) - The requested Processing object.

#### Response Example
```json
{
  "processing": {
    "id": "string",
    "name": "string",
    "creation_date": "string",
    "last_update": "string"
  }
}
```
```

--------------------------------

### Get Model by Name API

Source: https://documentation.picsellia.com/reference/client

Retrieve a model by its name.

```APIDOC
## GET /models?name={name}

### Description
Retrieve a model by its name.

### Method
GET

### Endpoint
/models?name={name}

### Parameters
#### Query Parameters
- **name** (str) - Required - The name of the model you are looking for.

### Response
#### Success Response (200)
- **model** (Model) - The requested Model object.

#### Response Example
{
  "id": "uuid-of-model",
  "name": "foo_model",
  "created_at": "timestamp"
}
```

--------------------------------

### Client Initialization

Source: https://documentation.picsellia.com/reference/client

Initialize the Picsellia SDK Client to interact with Picsellia services. Requires an API token.

```APIDOC
## Client Initialization

### Description
Initialize the Picsellia SDK Client to communicate with Picsellia services. An API token is required, which can be found in your Profile Settings on the web platform. The client allows you to retrieve and manipulate data objects within the platform.

### Arguments
* **api_token** (str, optional): Your API token. Defaults to None, in which case the client will attempt to find it in the environment variable 'PICSELLIA_TOKEN'.
* **organization_id** (str or UUID, optional): Specify an organization to connect to by its ID. Defaults to None, connecting to your main organization.
* **organization_name** (str, optional): Specify an organization to connect to by its name. If `organization_id` is also provided, `organization_id` will take precedence. Defaults to None, connecting to your main organization.
* **host** (str, optional): Define a custom host for the platform. Defaults to 'https://app.picsellia.com'.
* **session** (requests.Session, optional): Allows you to provide your own `requests.Session` object for custom headers or proxies.

### Request Example
```python
from picsellia import Client

client = Client(api_token="a0c9f3bbf7e8bc175494fc44bfc6f89aae3eb9d0")
```

### Properties
* **id** (Organization UUID): The UUID of the organization connected with this client.
* **name** (str): The name of the organization connected with this client.
```

--------------------------------

### Build and Push Docker Image

Source: https://documentation.picsellia.com/docs/part-4-make-your-modelversion-trainable

These bash commands are used to build a Docker image from a Dockerfile located in the current directory ('.') with a specified Dockerfile path ('your-custom-image/Dockerfile') and tag it with a registry name, image name, and version ('myregistry/my-image:1.0'). After building, the image is pushed to the Docker registry.

```bash
docker build . -f your-custom-image/Dockerfile -t myregistry/my-image:1.0
docker push myregistry/my-image:1.0
```

--------------------------------

### ModelVersion - Get File

Source: https://documentation.picsellia.com/reference/modelversion

Retrieves a specific ModelFile by its name from a ModelVersion.

```APIDOC
## GET /modelversions/{id}/files/{filename}

### Description
Retrieves a specific ModelFile by its name from a ModelVersion.

### Method
GET

### Endpoint
/modelversions/{id}/files/{filename}

### Parameters
#### Path Parameters
- **id** (string) - Required - The unique identifier of the ModelVersion.
- **filename** (string) - Required - The name of the ModelFile to retrieve.

### Response
#### Success Response (200)
- **file** (ModelFile) - The requested ModelFile object.

#### Response Example
```json
{
  "file": {
    "name": "model-latest",
    "content": "...file content..."
  }
}
```
```

--------------------------------

### Main Script Initialization with Picsellia Client

Source: https://documentation.picsellia.com/docs/part-4-make-your-modelversion-trainable

This Python script demonstrates how to initialize the Picsellia client by calling the `get_experiment` function and then listing attached dataset versions. It imports necessary libraries from TensorFlow, Picsellia, and standard Python modules. Ensure `utils.py` containing `get_experiment` is in the same directory or accessible.

```python
from tensorflow.keras.applications import EfficientNetB0
from tensorflow.keras.models import Sequential
from tensorflow.keras import layers
from tensorflow import keras
import tensorflow as tf
import tensorflow_datasets as tfds
from picsellia import Client, Experiment
from picsellia.types.enums import LogType
import os
import numpy as np 
from utils import get_experiment

print("Initializing Picsellia Client 🥑")

experiment = get_experiment()

datasets = experiment.list_attached_dataset_versions()

# The rest of your script
# ...
```

--------------------------------

### ModelVersion - Get Tags

Source: https://documentation.picsellia.com/reference/modelversion

Retrieves all tags associated with a specific ModelVersion.

```APIDOC
## GET /modelversions/{id}/tags

### Description
Retrieves all tags associated with a specific ModelVersion.

### Method
GET

### Endpoint
/modelversions/{id}/tags

### Parameters
#### Path Parameters
- **id** (string) - Required - The unique identifier of the ModelVersion.

### Response
#### Success Response (200)
- **tags** (list[Tag]) - A list of Tag objects associated with the ModelVersion.

#### Response Example
```json
{
  "tags": [
    {"name": "tag1", "target": "DATA"},
    {"name": "tag2", "target": "DATA"}
  ]
}
```
```

--------------------------------

### Model - Get Tags

Source: https://documentation.picsellia.com/reference/model

Retrieves all tags associated with a specific model.

```APIDOC
## Model - Get Tags

### Description
Retrieve the tags of your model.

### Method
GET

### Endpoint
/websites/picsellia/models/{model_id}/tags

### Parameters
#### Path Parameters
- **model_id** (string) - Required - The ID of the model whose tags are to be retrieved.

### Response
#### Success Response (200)
- **tags** (list[Tag]) - A list of Tag objects associated with the model.

#### Response Example
```json
[
  {
    "id": "tag-abc",
    "name": "my-model-tag",
    "target": "MODEL"
  }
]
```
```

--------------------------------

### Initialize Picsellia Client

Source: https://documentation.picsellia.com/reference/client

Initializes the Picsellia SDK Client with an API token. The token can be provided directly or will be read from the environment variable 'PICSELLIA_TOKEN'. Optional arguments allow specifying organization details and a custom host.

```python
client = Client(api_token="a0c9f3bbf7e8bc175494fc44bfc6f89aae3eb9d0")
```

--------------------------------

### Start Fast Training

Source: https://documentation.picsellia.com/reference/datasetversion

Initiates a fast training process using the current dataset as input data. An optional description can be provided.

```APIDOC
## POST /websites/picsellia/start_fast_training

### Description
This method will start a fast training with this dataset as input data.

### Method
POST

### Endpoint
/websites/picsellia/start_fast_training

### Parameters
#### Query Parameters
- **description** (str) - Optional - A description for the fast training job.

### Response
#### Success Response (200)
- **fast_training** (FastTraining) - A FastTraining object representing the initiated training.
```

--------------------------------

### Model Context - Get Experiment

Source: https://documentation.picsellia.com/reference/modelcontext

Retrieve the source experiment associated with this model context.

```APIDOC
## GET /websites/picsellia/modelcontext/experiment

### Description
Retrieve the source experiment of this context. It will raise an exception if this context has no experiment source.

### Method
GET

### Endpoint
/websites/picsellia/modelcontext/experiment

### Parameters
#### Query Parameters
None

### Request Example
None

### Response
#### Success Response (200)
- **experiment** (Object) - An Experiment object.

#### Response Example
{
  "id": "a1b2c3d4-e5f6-7890-1234-567890abcdef",
  "name": "My Experiment"
}
```

--------------------------------

### Create Model

Source: https://documentation.picsellia.com/reference/client

Creates a new model within the organization. Requires a name and optionally accepts a description.

```python
model = client.create_model(name="foo_model", description="A brand new model!")
```

--------------------------------

### Initialize Picsellia Client and Get Project/Experiment (Python)

Source: https://documentation.picsellia.com/docs/part-2-initializing-picsellia-connection-retrieve-the-experiment

This Python code snippet demonstrates how to initialize the Picsellia client using your API token, organization name, and host. It then retrieves a specific project and experiment by their names, preparing them for further use in your script. Ensure you replace placeholder values with your actual credentials.

```python
from picsellia import Client

client = Client(
  api_token="XXXXXXX", 
  organization_name="XXXX",
  host="https://app.picsellia.com"
)
project = client.get_project("documentation-project")
experiment = project.get_experiment("exp-0-documentation")
```

--------------------------------

### Get Model by ID

Source: https://documentation.picsellia.com/reference

Retrieve a specific model using its unique identifier.

```APIDOC
## GET /models/{id}

### Description
Retrieve a model by its id.

### Method
GET

### Endpoint
/models/{id}

### Parameters
#### Path Parameters
- **id** (str or UUID) - Required - The unique identifier of the model to retrieve.

### Response
#### Success Response (200)
- **model** (Model) - The requested Model object.

### Response Example
```json
{
  "id": "d8fae655-5c34-4a0a-a59a-e49c89f20998",
  "name": "My Model",
  "description": "A model for object detection"
}
```
```

--------------------------------

### Initialize Picsellia SDK Client

Source: https://documentation.picsellia.com/reference

Instantiate the Picsellia SDK Client to interact with Picsellia services. Requires an API token, which can also be set via an environment variable. Optional arguments allow specifying organization and custom host.

```python
client = Client(api_token="a0c9f3bbf7e8bc175494fc44сная3eb9d0")
```

--------------------------------

### Upload Instructions File for Campaign - Python

Source: https://documentation.picsellia.com/reference/annotationcampaign

Uploads an instructions file for an annotation campaign. The provided path should point to the local file to be uploaded.

```python
foo_campaign.upload_instructions_file("/path/to/file.pdf")
```

--------------------------------

### Get Asset Tag

Source: https://documentation.picsellia.com/reference/datasetversion

Retrieves an asset tag used in this dataset version by its name.

```APIDOC
## GET /websites/picsellia/get_asset_tag

### Description
Retrieve an asset tag used in this dataset version by its name.

### Method
GET

### Endpoint
/websites/picsellia/get_asset_tag

### Parameters
#### Query Parameters
- **name** (str) - Required - Name of the tag you're looking for

### Response
#### Success Response (200)
- **Tag** - The Tag object.

#### Response Example
```json
{
  "id": "tag_id_3",
  "name": "dog"
}
```
```

--------------------------------

### Get Job By ID

Source: https://documentation.picsellia.com/reference/client

Retrieves a specific Job object using its unique ID.

```APIDOC
## Get Job By ID

### Description
Get a [Job](job) from its id.

### Method
GET

### Endpoint
/jobs/{id}

### Parameters
#### Path Parameters
- **id** (UUID|str) - Required - The unique identifier of the job.

### Request Example
```python
job = client.get_job_by_id(
    id="YOUR JOB ID"
)
```

### Response
#### Success Response (200)
- **job** (Job) - The Job object corresponding to the provided ID.

#### Response Example
```json
{
  "id": "job_id_123",
  "name": "Example Job",
  "status": "COMPLETED",
  "created_at": "2023-10-27T10:00:00Z"
}
```
```

--------------------------------

### Get Model by ID API

Source: https://documentation.picsellia.com/reference/client

Retrieve a specific model using its unique identifier.

```APIDOC
## GET /models/{id}

### Description
Retrieve a model by its id.

### Method
GET

### Endpoint
/models/{id}

### Parameters
#### Path Parameters
- **id** (str or UUID) - Required - The ID of the model to retrieve.

### Response
#### Success Response (200)
- **model** (Model) - The requested Model object.

#### Response Example
{
  "id": "d8fae655-5c34-4a0a-a59a-e49c89f20998",
  "name": "my_model",
  "created_at": "timestamp"
}
```

--------------------------------

### Start Fast Training in Python

Source: https://documentation.picsellia.com/reference/datasetversion

Initiates a fast training process using the current dataset as input. An optional description can be provided for the training job.

```Python
start_fast_training(
   description: (str|None) = None
)
```

--------------------------------

### Create Deployment Tag

Source: https://documentation.picsellia.com/reference/client

Creates a tag that can be applied to Deployment objects. The tag is identified by its name.

```python
tag = client.create_deployment_tag("operation")
```

--------------------------------

### Get or Create Data Tag

Source: https://documentation.picsellia.com/reference/datalake

Retrieves a data tag by its name. If the tag does not exist, it is created.

```APIDOC
## POST /websites/picsellia/data/tags/get_or_create

### Description
Retrieves a data tag by its name. If the tag does not exist, it is created.

### Method
POST

### Endpoint
/websites/picsellia/data/tags/get_or_create

### Parameters
#### Request Body
- **name** (str) - Required - Name of the tag to retrieve or create.

### Request Example
```json
{
  "name": "new_tag"
}
```

### Response
#### Success Response (200)
- **Tag** (object) - A Tag object representing the retrieved or newly created tag.

#### Response Example
{
  "example": "Tag object"
}
```

--------------------------------

### ModelVersion - Get Context

Source: https://documentation.picsellia.com/reference/modelversion

Retrieves the ModelContext associated with a ModelVersion, which can be used to access further information and functionalities.

```APIDOC
## GET /modelversions/{id}/context

### Description
Retrieves the ModelContext associated with a ModelVersion, which can be used to access further information and functionalities.

### Method
GET

### Endpoint
/modelversions/{id}/context

### Parameters
#### Path Parameters
- **id** (string) - Required - The unique identifier of the ModelVersion.

### Response
#### Success Response (200)
- **context** (ModelContext) - The ModelContext object associated with the ModelVersion.

#### Response Example
```json
{
  "context": {
    "info": "Model context information"
  }
}
```
```

--------------------------------

### Create Deployment using Picsellia API

Source: https://documentation.picsellia.com/reference

Creates a Deployment from a specified ModelVersion. This enables monitoring and potentially a hosted endpoint. Optional parameters include shadow model version, name, minimum threshold, target datalake, and serving disablement.

```Python
def create_deployment(
   model_version: ModelVersion, shadow_model_version: (ModelVersion|None) = None,
   name: (str|None) = None, min_threshold: (float|None) = None,
   target_datalake: (Datalake|None) = None, disable_serving: (bool|None) = False
):
    """Create a Deployment from a model.
    This method allows you to create a Deployment on Picsellia. You will then have access to the monitoring dashboard and eventually a hosted endpoint.
    """
    pass

# Example:
model_version = client.get_model_version_by_id('918351d2-3e96-4970-bb3b-420f33ded895')
deployment = client.create_deployment(model_version=model_version)
```

--------------------------------

### Set Training Data for Deployment (Python)

Source: https://documentation.picsellia.com/reference/deployment

Sets the training data reference for a deployment, enabling metric computation based on training data distribution in the Monitoring service. Requires a DatasetVersion object.

```python
deployment.set_training_data(dataset_version)
```

--------------------------------

### Model - Get Version

Source: https://documentation.picsellia.com/reference/model

Retrieves a specific version of a model using its version number or name.

```APIDOC
## Model - Get Version

### Description
Retrieve a version of a model from its version or its name.

### Method
GET

### Endpoint
/websites/picsellia/models/{model_id}/versions/{version_identifier}

### Parameters
#### Path Parameters
- **model_id** (string) - Required - The ID of the model.
- **version_identifier** (integer | string) - Required - The version number or name of the model version to retrieve.

### Response
#### Success Response (200)
- **id** (string) - The ID of the model version.
- **version** (integer) - The version number.
- **name** (string) - The name of the model version.
- **framework** (string) - The framework used for this version.
- **type** (string) - The inference type for this version.
- **description** (string) - The description of the model version.

#### Response Example
```json
{
  "id": "version-xyz",
  "version": 0,
  "name": "first-version",
  "framework": "TENSORFLOW",
  "type": "OBJECT_DETECTION",
  "description": "Initial version of the model."
}
```

--------------------------------

### Initialize Picsellia Client and Fetch Project Data

Source: https://documentation.picsellia.com/docs/evaluate-a-detection-model-copy

This script initializes the Picsellia client connection and retrieves project, experiment, and dataset information. It requires API token, organization name, and project/experiment/dataset names as parameters.

```python
from picsellia import Client 

client = Client(api_token, organization_name, host='https://app.picsellia.com')

project = client.get_project(name='Documentation Project')
experiment = project.get_experiment(name='my_experiment')

testing_dataset = experiment.get_dataset('test')
```

--------------------------------

### Get Processing Information

Source: https://documentation.picsellia.com/reference

Retrieves details about a specific processing job within the organization by its name.

```APIDOC
## GET /websites/picsellia/processing

### Description
Find a processing in this organization by its name.

### Method
GET

### Endpoint
/websites/picsellia/processing

### Parameters
#### Query Parameters
- **name** (str) - Required - The name of the processing job to find.

### Request Example
```
GET /websites/picsellia/processing?name=auto-tagging-dataset
```

### Response
#### Success Response (200)
- **Processing** (object) - A Processing object containing details of the found processing job.

#### Response Example
```json
{
  "id": "60d5ec49a1b2c3d4e5f6a7b8",
  "name": "auto-tagging-dataset",
  "status": "completed",
  "createdAt": "2023-10-27T10:00:00Z"
}
```
```

--------------------------------

### Experiment Launch and Artifacts

Source: https://documentation.picsellia.com/reference/experiment

Launch experiments on remote environments and download their artifacts.

```APIDOC
## POST /experiments/{experiment_id}/launch

### Description
Launch a job on a remote environment with this experiment. The remote environment must be set up prior to launching.

### Method
POST

### Endpoint
/experiments/{experiment_id}/launch

### Parameters
#### Request Body
- **gpus** (integer, optional) - Number of GPUs to use for the training. Defaults to 1.

### Request Example
```json
{
  "gpus": 2
}
```

### Response
#### Success Response (200)
- **job_id** (string) - The ID of the launched job.

#### Response Example
```json
{
  "job_id": "job_789"
}
```

## GET /experiments/{experiment_id}/artifacts

### Description
Download all artifacts from the experiment to the local directory.

### Method
GET

### Endpoint
/experiments/{experiment_id}/artifacts

### Parameters
#### Query Parameters
- **with_tree** (boolean, optional) - If true, the artifacts will be downloaded in a tree structure.

### Response
#### Success Response (200)
- **download_url** (string) - A URL to download the artifacts.

#### Response Example
```json
{
  "download_url": "https://example.com/artifacts.zip"
}
```
```

--------------------------------

### List Deployment Tags

Source: https://documentation.picsellia.com/reference/client

Lists all available Deployment tags. These tags are applicable only to Deployment objects.

```APIDOC
## List Deployment Tags

### Description
List all Deployment tags, usable only on Deployment Version objects.

### Method
GET

### Endpoint
/deployments/tags

### Parameters
#### Query Parameters
None

### Request Example
```python
tags = client.list_deployment_tags()
```

### Response
#### Success Response (200)
- **tags** (list[Tag]) - A list of Tag objects.

#### Response Example
```json
[
  {
    "id": "tag_id_1",
    "name": "tag_name_1",
    "color": "#FF0000"
  },
  {
    "id": "tag_id_2",
    "name": "tag_name_2",
    "color": "#00FF00"
  }
]
```
```