crambl / mtgo-collection-manager Goto Github PK

Creating a symlink to the mtgo-collection-manager executable and provoking an update of the full collection data, results in the appdata being written to the location of the symlink instead of adjacent (in the same directory) as the bin folder.

Relevant cpp weekly episodes:

• How (and Why) To Write Code That Avoids std::move
• The Optimal Way To Return From A Function

make it ignore all build directories

There's an info section with collection info, a similar one could be made but with meta data such as:

• Goatbots prices last updated
• Scryfall prices last updated
• Next set release date
• (dek-file added moved from collection info to this meta-data section
• Next available update?

Show:

• When was the tradelist set (date)
• How many cards (unique and total)
• Total value
• Rarity distribution?
• Max value?
• Other?

directory with data from each time the collection data was updated, and/or a single file that continuously adds information (would save space, but be more complicated and less modular).

Potential space issues could be resolved with compression.

It could also be user configurable e.g. limit history to a duration of 3 months.

New releases (including near future ones) can be tracked from https://scryfall.com/docs/api/sets/all.

Not entirely sure what features are coupled with this, but something should probably happen around when a new release is out.

Several viable options: Installable as static library, precompiled binary, shared library with a config describing dependencies...

Initial goal

• Functionality for the preprocessor to aggregate all the price history data into a single CSV file and write it to stdout

Make a file saving the last time it was done to prevent doing all the work every time? Or rename the file each time time with the timestamp.
If a "running" aggregate file is used, compress all files that have been aggregated previously?

• MTGO GUI receives this and does something meaningful with it

Using CSV, format and parsing implemted in #91

#94

• Functionality to read all previously processed price data JSON-files from a directory, in chronological order and aggregate is into a price history class containing historic data of user collection quantity and price data for each card.
• Functionality to deserialize and serialize it to CSV
• Functionality to read a previously aggregated CSV and append new price data and save it once again

Next step

• Functionality to continuously archive (zip) all files added to the price history CSV
• Add the behaviour to the run update scenario of mtgo_preprocessor

Final step

• Print the CSV to stdout and implement decoding etc. behaviour in the GUI

Toggle and remember at startup (display tooltip)

Explore including MTGO Getter and MTGO Preprocessor with include_bytes! and check for them at startup, writing them out if they're not present.

Additionally, add a build script for MTGO GUI that builds the binaries to facilitate the include_bytes!

Seems promising

https://crates.io/crates/fltk

Add a tab (or window?) with a trade tool that lets you enter cards and see a price comparison and a total cost for acquiring those cards.

Trade tool should have:

• Toggle option to just select cheapest version of each entered card
• When a card is selected -> popup where user has to select set/version of card from available sets for that card
• Box showing where to buy each cards to get the best prices
• Time showing how old the price data is for each vendor

Sort icons stick currently, make it so that it is only shown on the column that was sorted most recently.

https://en.cppreference.com/w/cpp/types/floating-point

fltk-rs/fltk-rs#1466

and add a timestamp. Load the saved full trade list at startup and display a tooltip with how old the full trade list is (duration since, and date added).

Treating warnings as errors is now disabled as dependencies generate warnings on certain OS/compiler combinations.

Currently it is read to a byte array and decoded with json.unmarshal. Especially for the giant JSON array download via the scryfall API there could be performance gains from decoding a stream instead of unmarshalling when everything is in-memory.

Window/tab for entering cards that might not be in your collection but you want to monitor their prices.

Could reuse many of the components of the trade tool (see #88)

Currently using std::unordered_map which is ~5 times slower than Abseil's flat_hash_map . Change the map to Abseil's or another even faster one (perfect hashing should be achievable with at least some of the data).

https://www.boost.org/doc/libs/1_83_0/boost/implicit_cast.hpp