Merge branch 'main' into task/out-of-kb-questions-gsk-3372

Author: mattbit

docs/_templates/base.html

@@ -61,15 +61,15 @@
     {%- block htmltitle -%}
     {% if not docstitle %}
     <title>{{ title|striptags|e }}</title>
-    <meta content="Learn more about {{ title|striptags|e }} | Open source Python ML library to test and debug AI models."
+    <meta content="Learn more about Giskard {{ title|striptags|e }} | The Testing platform for AI models."
           name="description">
     {% elif pagename == master_doc %}
     <title>{{ docstitle|striptags|e }}</title>
-    <meta content="Learn more about Giskard | Open source Python ML library to test and debug AI models."
+    <meta content="Giskard Documentation | The Testing platform for AI models."
           name="description">
     {% else %}
     <title>{{ title|striptags|e }} - {{ docstitle|striptags|e }}</title>
-    <meta content="Learn more about {{ title|striptags|e }} | Open source Python ML library to test and debug AI models."
+    <meta content="Learn more about Giskard {{ title|striptags|e }} | The Testing platform for AI models."
           name="description">
     {% endif %}
     {%- endblock -%}
@@ -128,4 +128,4 @@
 
 {%- endblock scripts -%}
 </body>
-</html>
+</html>

docs/cli/ngrok/index.rst

@@ -31,7 +31,13 @@ Now you can run :code:`giskard hub expose --ngrok-token <ngrok_API_key>` which s
     token=...
     client = giskard.GiskardClient("<ngrok_external_server_link>", token)
 
-    # To run your model with the Giskard Hub, execute these three lines on Google Colab:
+    # To run your model with the current Python environment on Google Colab, execute these lines:
 
     %env GSK_API_KEY=...
-    !giskard worker start -d -u <ngrok_external_server_link>
+    !giskard worker start -d -u <ngrok_external_server_link> --name <your-project-worker-id>
+
+    # To let Giskard Hub run your model in a managed Python environment, execute these lines:
+
+    %env GSK_API_KEY=...
+    !giskard worker start -s -u <ngrok_external_server_link> --name <your-project-worker-id>
+

docs/getting_started/index.md

@@ -1,33 +1,30 @@
 # Why Giskard?
 
-Giskard is an open-source **AI quality management system** dedicated to ML models.
+Giskard is a **holistic Testing platform for AI models** to control all 3 types of AI risks: Quality, Security & Compliance.
 
 It addresses the following challenges in AI testing:
 
-* Edge cases in AI are **domain-specific** and often seemingly **infinite**
-* The AI development process is an experimental, **trial-and-error** process where quality KPIs are multi-dimensional
-* AI regulation and standardization necessitate that data scientists write **extensive documentation** and quality reports
+* Edge cases in AI are **domain-specific** and often seemingly **infinite**.
+* The AI development process is an experimental, **trial-and-error** process where quality KPIs are multi-dimensional.
+* Generative AI introduces new **security vulnerabilities** which requires constant vigilance and adversarial red-teaming.
+* AI compliance with new regulations necessitate that data scientists write **extensive documentation**.
 
-Giskard provides a suite of tools for **scanning**, **testing**, **debugging**, and **monitoring** all AI models, from tabular to LLMs. This enables AI engineers to:
-1. Save time during the **test-writing process**
-2. Enhance the coverage rate of the testing process through **domain-specific tests**
-3. Strengthen the CI/CD pipeline by **automating** test execution
-4. Save time in writing **quality metrics** and reports
+Giskard provides a platform for testing all AI models, from tabular ML to LLMs. This enables AI teams to:
+1. **Reduce AI risks** by enhancing the test coverage on quality & security dimensions.
+2. **Save time** by automating testing, evaluation and debugging processes.
+3. **Automate compliance** with the EU AI Act and upcoming AI regulations & standards.
 
-Giskard currently offers 3 tools for AI quality management: the **Giskard open-source Python library**, the **Giskard Quality
-Assurance Hub** and the **LLM Monitoring platform (LLMon)**.
 
-## The Giskard open source Python Library
+## Giskard Library (open-source)
 
-An **open-source** library to scan your AI models for vulnerabilities and generate test suites automatically to aid in
-the Quality Assurance process of ML models and LLMs.
+An **open-source** library to scan your AI models for vulnerabilities and generate test suites automatically to aid in the Quality & Security evaluation process of ML models and LLMs.
 
-Testing Machine Learning applications can be tedious. Since ML models depend on data, testing scenarios depend on
-**domain specificities** and are often **infinite**.
+Testing Machine Learning applications can be tedious. Since AI models depend on data, quality testing scenarios depend on
+**domain specificities** and are often **infinite**. Besides, detecting security vulnerabilities on LLM applications requires specialized knowledge that most AI teams don't possess.
 
-Created by ML engineers for ML engineers, `giskard` enables you to:
+To help you solve these challenges, Giskard library helps to:
 
-- **Scan your model to find dozens of hidden vulnerabilities**: The `giskard` scan automatically detects vulnerabilities
+- **Scan your model to find hidden vulnerabilitie automatically**: The `giskard` scan automatically detects vulnerabilities
 such as performance bias, hallucination, prompt injection, data leakage, spurious correlation, overconfidence, etc.
   <br><br>
   <iframe src="https://htmlpreview.github.io/?https://gist.githubusercontent.com/AbSsEnT/a67354621807f3c3a332fca7d8b9a5c8/raw/588f027dc6b14c88c7393c50ff3086fe1122e2e9/LLM_QA_IPCC_scan_report.html" width="700" height="400"></iframe>
@@ -39,7 +36,7 @@ vulnerabilities detected in the scan.
   <img src="../assets/test_suite_scan_llm.png" width="500">
 
 
-- **Integrate and automate** the quality testing of AI models in **CI/CD** processes by leveraging native `giskard` integrations.
+- **Integrate and automate** testing of AI models in **CI/CD** pipelines by leveraging native `giskard` integrations.
   <br><br>
   <img src="../assets/gh_discussion.png" width="650">
 
@@ -48,8 +45,7 @@ Get started **now** with our [quickstart notebooks](../getting_started/quickstar
 
 ## Giskard Hub
 
-An enterprise AI quality management platform for ML engineers, domain experts and AI Quality Assurance teams to manage
-all their AI model quality testing and debugging activities in a centralized hub.
+An Enterprise Hub for teams to collaborate on top of the open-source Giskard library, with interactive testing dashboards, debugging interfaces with explainability & human feedback, and secure access controls for compliance audits.
 
 - 🔍 **Debug** your issues by inspecting the failing examples of your tests (⬇️ see below the DEBUG button)
   <br><br>
@@ -63,23 +59,10 @@ all their AI model quality testing and debugging activities in a centralized hub
   <br><br>
   ![](../assets/push.png)
 
-- 💬 **Collect business feedback** and **share your model results** with data scientists, QA teams and decision makers.
+- 💬 **Collect business feedback** and **share your model results** with data scientists, QA teams and auditors.
   <br><br>
   ![](../assets/credit_scoring_comment.png)
 
 
 Get started **now** with our [demo HuggingFace Space](https://huggingface.co/spaces/giskardai/giskard) or
 by following [installation instructions](../getting_started/quickstart/index.md)! 🐢
-
-<br>
-
-## LLM Monitoring (LLMon)
-
-A SaaS platform to detect AI Safety risks in your deployed LLM applications. From hallucinations to incorrect responses, toxicity and many more metrics.
-
-Get access to insights on your LLM app performance in 2 lines of code, giving you the ability to evaluate LLM quality in real-time.
-
-  ![](../assets/llm_monitoring_dashboard.gif)
-
-
-Get started **now** by [signing up for our beta](https://www.giskard.ai/products/llmon)! 🍋

docs/giskard_hub/installation_hub/index.md

@@ -13,7 +13,7 @@ The Giskard Hub is the app adapted for an enterprise use of Giskard. Extending t
 You have 3 ways to install the Hub:
 * **Hugging Face Space installation**: This is adapted for an **easy installation** in the cloud for prototyping purposes. If you don't want to upload your own model and just want to check some Giskard demo projects, the HF public space is perfect for you.
 * **On-premise installation**: This is adapted if your data and model are **private** and you don't have the possibility to use the cloud (for instance, because of privacy and connectivity issues).
-* **Private cloud**: This is adapted if you can easily run a **Cloud instance** by your favourite Cloud provider (AWS, GCP or Azure) and want to easily use Giskard collaborative features (collect feedback from business, share results, etc.). Make sure that you have the rights to open ports of your Cloud machine because Giskard needs to open a connection with an ML Worker running on your Python environment.
+* **Private cloud**: This is adapted if you can easily run a **Cloud instance** by your favourite Cloud provider (AWS, GCP or Azure) and want to easily use Giskard collaborative features (collect feedback from business, share results, etc.). Make sure that you have the rights to open ports of your Cloud machine, if you want to connect your own ML Worker to the Giskard Hub.
 
 ```{toctree}
 :caption: Table of Contents

docs/giskard_hub/installation_hub/install_cloud/install_aws/index.md

@@ -44,72 +44,8 @@ You can stop the instance and restart it when you need to save AWS compute costs
 
 ## 5. Start the ML worker
 
-Giskard executes your model using a worker that runs the model directly in **your Python environment**, with all the dependencies required by your model. You can either execute the ML worker:
+Giskard executes your model using an ML worker that runs the model. The worker is created along with your project, using the dependencies in your current environment. You can start the worker on Giskard Hub, if it is not started automatically.
 
-- From your **local notebook** within the kernel that contains all the dependencies of your model
-- From **Google Colab** within the kernel that contains all the dependencies of your model
-- Or from **your terminal** within the Python environment that contains all the dependencies of your model
-
-:::{note}
-If you plan to use LLM-assisted tests or transformations, don’t forget to set the ``OPENAI_API_KEY`` environment
-variable before starting the Giskard worker.
-:::
-
-:::::::{tab-set}
-::::::{tab-item} From your local notebook
-
-To start the ML worker from your notebook, run the following code in your notebook:
-
-```
-!giskard worker start -d -k YOUR_KEY -u http://<your IP address>:19000/
-```
-
-The API Access Key (`YOUR_KEY`) can be found in the Settings tab of the Giskard Hub.
-
-
-> ### ⚠️ Warning
-> To see the available commands of the worker, you can execute:
->```
->!giskard worker --help
->```
+For advanced and flexible usages, please check [our doc for ML worker](../../mlworker/index.md).
 
 You're all set to try Giskard in action. Upload your first model, dataset or test suite by following the [upload an object](../../../upload/index.md) page.
-
-::::::
-::::::{tab-item} From Colab notebook
-
-To start the ML worker from your Colab notebook, run in your Colab cell:
-
-```
-!giskard worker start -d -k YOUR_KEY -u http://<your IP address>:19000/
-```
-The API Access Key (`YOUR_KEY`) can be found in the Settings tab of the Giskard Hub.
-
-> ### ⚠️ Warning
-> To see the available commands of the worker, you can execute:
->```
->!giskard worker --help
->```
-
-You're all set to try Giskard in action. Upload your first model, dataset or test suite by following the [upload an object](../../../upload/index.md) page.
-
-::::::
-::::::{tab-item} From your terminal
-
-* Run the following command **within the Python environment that contains all the dependencies of your model**:
-
-```
-giskard worker start -k YOUR_KEY -u http://<your IP address>:19000/
-```
-The API Access Key (`YOUR_KEY`) can be found in the Settings tab of the Giskard Hub.
-
-> ### ⚠️ Warning
-> To see the available commands of the worker, you can execute:
->```
->!giskard worker --help
->```
-
-You're all set to try Giskard in action. Upload your first model, dataset or test suite by following the [upload an object](../../../upload/index.md) page.
-
-::::::
-:::::::

docs/giskard_hub/installation_hub/install_cloud/install_azure/index.md

@@ -57,72 +57,8 @@ You can stop the instance and restart it when you need to save your Azure comput
 
 ## 4. Start the ML worker
 
-Giskard executes your model using a worker that runs the model directly in **your Python environment**, with all the dependencies required by your model. You can either execute the ML worker:
+Giskard executes your model using an ML worker that runs the model. The worker is created along with your project, using the dependencies in your current environment. You can start the worker on Giskard Hub, if it is not started automatically.
 
-- From your **local notebook** within the kernel that contains all the dependencies of your model
-- From **Google Colab** within the kernel that contains all the dependencies of your model
-- Or from **your terminal** within the Python environment that contains all the dependencies of your model
-
-:::{note}
-If you plan to use LLM-assisted tests or transformations, don’t forget to set the ``OPENAI_API_KEY`` environment
-variable before starting the Giskard worker.
-:::
-
-:::::::{tab-set}
-::::::{tab-item} From your local notebook
-
-To start the ML worker from your notebook, run the following code in your notebook:
-
-```
-!giskard worker start -d -k YOUR_KEY -u http://<your IP address>:19000/
-```
-
-The API Access Key (`YOUR_KEY`) can be found in the Settings tab of the Giskard Hub.
-
-
-> ### ⚠️ Warning
-> To see the available commands of the worker, you can execute:
->```
->!giskard worker --help
->```
-
-You're all set to try Giskard in action. Upload your first model, dataset or test suite by following the [upload an object](../../../upload/index.md) page.
-
-::::::
-::::::{tab-item} From Colab notebook
-
-To start the ML worker from your Colab notebook, run in your Colab cell:
-
-```
-!giskard worker start -d -k YOUR_KEY -u http://<your IP address>:19000/
-```
-The API Access Key (`YOUR_KEY`) can be found in the Settings tab of the Giskard Hub.
-
-> ### ⚠️ Warning
-> To see the available commands of the worker, you can execute:
->```
->!giskard worker --help
->```
-
-You're all set to try Giskard in action. Upload your first model, dataset or test suite by following the [upload an object](../../../upload/index.md) page.
-
-::::::
-::::::{tab-item} From your terminal
-
-* Run the following command **within the Python environment that contains all the dependencies of your model**:
-
-```
-giskard worker start -k YOUR_KEY -u http://<your IP address>:19000/
-```
-The API Access Key (`YOUR_KEY`) can be found in the Settings tab of the Giskard Hub.
-
-> ### ⚠️ Warning
-> To see the available commands of the worker, you can execute:
->```
->!giskard worker --help
->```
+For advanced and flexible usages, please check [our doc for ML worker](../../mlworker/index.md).
 
 You're all set to try Giskard in action. Upload your first model, dataset or test suite by following the [upload an object](../../../upload/index.md) page.
-
-::::::
-:::::::

docs/giskard_hub/installation_hub/install_cloud/install_gcp/index.md

@@ -62,76 +62,10 @@ You can stop the instance and restart it when you need to save your GCP compute
  `giskard hub start`
 :::
 
-
-
 ## 4. Start the ML worker
 
-Giskard executes your model using a worker that runs the model directly in **your Python environment**, with all the dependencies required by your model. You can either execute the ML worker:
-
-- From your **local notebook** within the kernel that contains all the dependencies of your model
-- From **Google Colab** within the kernel that contains all the dependencies of your model
-- Or from **your terminal** within the Python environment that contains all the dependencies of your model
-
-:::{note}
-If you plan to use LLM-assisted tests or transformations, don’t forget to set the ``OPENAI_API_KEY`` environment
-variable before starting the Giskard worker.
-:::
-
-:::::::{tab-set}
-::::::{tab-item} From your local notebook
-
-To start the ML worker from your notebook, run the following code in your notebook:
-
-```
-!giskard worker start -d -k YOUR_KEY -u http://<your IP address>:19000/
-```
-
-The API Access Key (`YOUR_KEY`) can be found in the Settings tab of the Giskard Hub.
-
-
-> ### ⚠️ Warning
-> To see the available commands of the worker, you can execute:
->```
->!giskard worker --help
->```
-
-You're all set to try Giskard in action. Upload your first model, dataset or test suite by following the [upload an object](../../../upload/index.md) page.
-
-::::::
-::::::{tab-item} From Colab notebook
-
-To start the ML worker from your Colab notebook, run in your Colab cell:
-
-```
-!giskard worker start -d -k YOUR_KEY -u http://<your IP address>:19000/
-```
-The API Access Key (`YOUR_KEY`) can be found in the Settings tab of the Giskard Hub.
+Giskard executes your model using an ML worker that runs the model. The worker is created along with your project, using the dependencies in your current environment. You can start the worker on Giskard Hub, if it is not started automatically.
 
-> ### ⚠️ Warning
-> To see the available commands of the worker, you can execute:
->```
->!giskard worker --help
->```
+For advanced and flexible usages, please check [our doc for ML worker](../../mlworker/index.md).
 
 You're all set to try Giskard in action. Upload your first model, dataset or test suite by following the [upload an object](../../../upload/index.md) page.
-
-::::::
-::::::{tab-item} From your terminal
-
-* Run the following command **within the Python environment that contains all the dependencies of your model**:
-
-```
-giskard worker start -k YOUR_KEY -u http://<your IP address>:19000/
-```
-The API Access Key (`YOUR_KEY`) can be found in the Settings tab of the Giskard Hub.
-
-> ### ⚠️ Warning
-> To see the available commands of the worker, you can execute:
->```
->!giskard worker --help
->```
-
-You're all set to try Giskard in action. Upload your first model, dataset or test suite by following the [upload an object](../../../upload/index.md) page.
-
-::::::
-:::::::