Reduced dimensionality of device state to only include the name of the capabilities (not their values).

Given a test name returns a list of challenges the test poses

Get text embeddings

Get available tools to agent (at given level of hierchy)

Get a unique dirname.

Gets the version.

Check if device has refresh capability If it does, you should use this capability before reading attributes, or they might n

Identify devices based on command string The format of the command string should be the same as that described in the DeviceDisambig

Get list of available tools, at every hierchy level of the coordinator agent

If the save folder contains saved instructions, load these instructions

Convert output of dump_ordered_dict_recur to OrderedDict

Load human-written instructions

Main function to generate the memory bank for multiple users. This is done as follows: 1- If a memory bank already exists, add manual mem

Given a prompt, sends a request to ChatGPT

Overrides the LLM config for the tools based on the coordinator config

Check whether the polling_fn has found anything

Auto-populate the DB. You only need to run this when devices are updated.

Prepare documents from the users' history for embedding and vector storage

Beatifully print the daily summaries

Beatifully print the global profiles

Helper to pretty print config to terminal

Read the latest json file in a folder

Replace requests with fake requests in a code string.

Run the LLM on the given prompt and input.

Sample n machine instructions from a list of machine instructions.

Saves the memory into a json file

Save config to output_dir directory

Either save an item to its own file or concatenate it with all dataset items in a single file.

Applies a fuzzy matching based on the number of matching characters. Necessary because LLMs sometimes make mistakes copying the IDs.

Selects instructions based on how diverse they are

Sets the global test id.

Tool-specific setup

Setup the memory retrieval chain

Wrapper function to visualize graph in a loop from given logs path at given loop frequency

Live visualization of the log file as a decision tree

Convert a json to a dataframe

Save self to disk as JSON.

Convert JSON objects (i.e. object made of lists and dictionaries) to use OrderedDict with sorted key instead of regular dictionaries. Th

Extract daily user preferences based on daily interactions

Update the user memory.

Validates initial state json files. Assumes filenames are of format `h#s#.json`

Live visualization of the log file as a decision tree

What's scheduled for this provider? Args: provider_string now_utc: UTC time for which to check the schedule.

Clean up processes if server is deleted.

Destructor to cleanup when done.

Initialize callback handler.

Initialize DatasetWriter.

Add a condition to be monitored by the server.

Function that is called to check if any triggers have been detected by pollers.

Allows users to post triggers directly to server for testing.

Wipe out all conditions.