fixess 0.0.6

Description

The Fixess Predictive Database returns an AI prediction instead of raising a KeyError for missing keys. This is a Python library that interacts with the fixess cloud database to provide cloud GPU accelerated predictions for missing data.

Getting Started

Get an X-RapidAPI-Key from https://rapidapi.com/fixessgithub/api/fixess To start using this package, you need to pass in your Rapid API headers (API Key) to create a fixess dictionary client:

import fixess
client = fixess.Fixess(api_key=YOUR_API_KEY)

Replace YOUR_API_KEY with your actual Rapid API Key.

Usage

Once you have created a client object, you can use it like a Python dictionary. For example:

value = client['missing-key']
print(value)

# Or with a more complex key
value = client[{'key' : 'value'}]
print(value)

In the above code snippet, missing-key and {'key' : 'value'} are keys absent in the database. Still, instead of raising a KeyError, the Fixess Predictive Database provides a sophisticated AI prediction.

The fixess API will automatically serialize (pickle) your dictionary keys, allowing you to use complex types as keys.

Please make sure to use the client object like a python dictionary, as shown in the example above.

Enjoy using Fixess Predictive Database!

For more information, visit our official documentation or submit an issue on our Github page.

Warning

An important note for users considering storing keys from multiple users in a single Fixess predictive database:

Each Fixess predictive database is backed by a single AI model. This has implications for the privacy and isolation of data.

When storing sensitive data from multiple users in a single database, it's possible for the AI model to share information inferred from this data between users during the prediction process. This is a byproduct of how the AI model learns patterns across all the data it has been given access to.

Please note, this does not imply that data will be explicitly shared between users nor does it mean that data will be shared between distinct databases or between distinct AI models.

While the raw data itself always remains strictly isolated (the database will never show one user's data to another user), the AI model's predictions can inevitably be influenced by the data from all users.

Therefore, if stringent data isolation is a requirement, we recommend against storing sensitive data from different users in the same predictive database. Consider creating separate databases for each user or thoroughly anonymizing data prior to insertion.

Always be attentive towards user privacy and respectful of all relevant data protection laws and best practices when using Fixess predictive databases.

License

MIT License