Link Sync

Purpose

The goal of this project is to sync gcode files from a local "master" directory to any number of Prusa printers on the local network running PrusaLink when they are idle, simultaneously.

Transfering files is tediously slow, so the idea is to sync several printers all at once. This is a powerful workflow for farm operators who may have a new model that replaces an older one and negates the need to send it individually to every printer, or, to manually fetch their USB stick/other to for updating at your local machine.

Also, by removing files that aren't in the local "master" directory, once off prints are automatically cleaned up from a printer's USB stick.

Theory of Operation

Locally, a "master" file structure is maintained containing all the GCode files required at the target machines. Also, configuration file in either YAML or JSON format is kept on the local machine that contains the connection details for each printer.

The program reads the configuration file, instantiates printer instances, then for the printers that aren't currently busy, a thread is spawned which is responsible for check what files should be deleted from the printer, or copied to the printer to bring it into sync with the local "master" file-structure.

These threads then operate simultaneously to bring all idle printers into sync at roughly the same time.

Note that PrusaLink running on the printer refuses to delete a file that is actively being printed. This is a good thing.

Being a command-line program, it would be easy to set up with cron to run every hour on the hour from 7pm to 7am (overnight) to ensure that all printers get synced/cleaned up from the previous day's run and ready to go for the next day.

Currently Implemented Features

Printers currently supported:

• Prusa MK4 running firmware 5.0.0 RC1

TODO:

• Prusa XL
• Prusa Mini
• Prusa MK3

Features:

• Add new files that were found in local source but not in remote storage.
• Delete excess files not found in local source but are in remote storage.
• Optionally operate on printers that are not idle.
• Replace stale files that have a modification date more recent (> 60 sec.) than the same file on the remote storage.

TODO:

• Optionally, copy files that exist on a printer that are unique back to the local filesystem.

Program Structure

classDiagram

    class FileNode {
        +child_nodes : Set[FileNode]
        +children : List[Dict] or None
        +display_name : str
        +full_display_path : str
        +full_short_path : str
        +is_dir : bool
        +m_timestamp : int or None
        +mtime_as_date : datetime
        +name : str
        +parent_node : FileNode or None
        +ro : bool
        +type : str
        #kwargs : Dict
        #short_name : str
    }

    class AbstractApi {
        +name : str
        +host : str
        +username : str
        +password : str
        +status_response()
        +files_response()
    }

    class ApiResponse {
        +status_code : str
        +payload : object or None
        +success : bool
    }

    class PrusaLinkApi {
        +name : str
        +host : str
        +username : str
        +password : str
        #auth
        #session
        +status_response() : ApiResponse
        +files_response() : ApiResponse
    }

    class Storage {
        +name : str
        +api_key : str
        +api : PrusaLinkApi
        +root_node : FileNode
        #build_storage()
        +create_remote_folder()
        +delete_file(): bool
        #get_root_node() : FileNode
        +gen_nodes() : Generator[FileNode]
        +get_node_for_display_path() : FileNode
        +get_node_for_short_path() : FileNode
        +get_shorter_path() : Path
        +reload()
        +upload_file() : bool
    }

    class Printer {
        +name : str
        +printer_type : str
        +storage : Storage
        +gen_missing_or_stale_files() : Generator[Tuple[Path, Path]]
        +get_excess_files() : Set[Path]
        +get_state() : State
    }

    Printer --* Storage : Composition
    Storage --o FileNode : Aggregates
    Storage --* PrusaLinkApi : Composition
    PrusaLinkApi --|> AbstractApi : Implements

    AbstractApi -- ApiResponse : Associates
    PrusaLinkApi -- ApiResponse : Associates

The __main__ module

There is certain UI-related work performed in the link_sync package's __main__ module including the reading of a configuration file to instantiate Printer instances. It is also here that a scan of the local "master" filesystem is performed before a thread-pool is created to operate on as many printers as there are local CPU cores at a time.

The primary entry point is the process() function which manages reading the local "master" file-structure and managing the pool of threads which all run a _sync() function.

The Printer class

This represents a single printer instance. The printer will instantiate a Storage instance on initialization, passing in all the credentials required for an API instance.

The Storage class

The Storage class manages its own PrusaLinkApi instance the collection of files that are all linked to the root_node of the Printer's storage. For example, on Prusa MK4 printers, this is an object with the name usb. From here, further FileObjects are linked as child_nodes, and so on representing the whole file-system.

Any given file references its parent in FileNode parent_node and directory ("FOLDER") objects contain their children as a set of FileNodes in child_nodes.

The FileNode class

This represents a single object on the remote storage. However, FileNode instances are not natively, but can be doubly linked together by logic in the Storage class.

The PrusaLinkApi class

This is a class that subclasses the abstract AbstractApi class. THe idea is that other types of APIs could also implement the required methods of the AbstractApi in the future (AbstractApi should probably be a Python Protocol instead).

The PrusaLinkApi implements the required methods specifically for PrusaLink 0.7.0 which is under fairly rapid development/evolution.

Although this will likely change in the near future, the only required methods are to handle requests to PrusaLink's files and status endpoints using various HTTP verbs (GET, PUT, and DELETE, specifically.)

The ApiResponse class

This is a small data class to hold an API response's status_code and deserialized JSON payload with some helper methods.

Copyright (c) 2023, Martin Koistinen