Every cell is re-created and re-drawn each time, which kinda ruins the whole point

Should dive a little bit deeper and find a way to avoid it. Probably by creating the cells once and then just updating their properties
BTW, that will also facilitate moving from two parallel data structures (1. cells 2. drawn cells) to one --> make everything easier. Composition will be stored in cell properties then

Current one only worked once and is kinda unintuitive anyway.
Simple weighted average seems to work just fine

Possible Future Features to Tackle (PFFT)

This list includes some features that are not in the "zone of proximal development" for now, and thus may never actually be implemented. Especially, given the nature of this project. However, they are cool and would theoretically make everything better

PFFT list:

• in-built statistics & graphs 📈📊
• csv/xml data input 📃
• csv data output
• performance upgrades, on/off/partial visuals⏱️
• beautiful, stunning, sexy visuals 🤩
• actual thermodynamics constants 🔥
• actual thermodynamics equations🌡️
• inverse problem solving 🔮😮
• some sort of interaction with the big boys (COMAGMAT💪 etc)
• different chamber shape than just a rectangle 🔷
• 3D 👀
• interaction with host rock beyond temperature 🌋
• web version 🕸️

• Maybe via number of cells travelled per step
• But preferably via different probabilities to sink on any given step for different crystals

matplotlib.pyplot colormap seem suboptimal for visualisation once ot becomes more complex.

Visualisation features reqiured:

• Continious color for liquid phase with different temperatures, discrete colors for different phases --> at least two different data fields should influence the appearece of the cell
• Maybe different symbols for different phases (dwarf-fortress style lol) --> customization beyond color required
• Positively have no idea how to use an array of cell class with pyplot :(

Alternatives

• PIL <-- can either save images or display using external software, not draw&update
• plotly <-- easy animation, but requires all data before drawng
• pygame <-- has in-built draw loop

Should support on-the-fly updates too

As of now:

• Phase composition is not implemented.
• The only crystalline phase is olivine.
• Crystallization of a given cell is determined by probability

How it should be:

• Liquid phase should have pre-determined composition (MgO, FeO, CaO, SiO2, Al2O3)
• Multiple different crystal phases should be able to form depending on current liquid composition
• Crystallization should turn liquid cell into crystal AND distribute unused components between the neighbors
• Conservation of mass should apply
• Instead of probability-based could be proximity-based: crystal always forms unless there already one forming nearby

Temperature, phases, their composition etc

The idea is logging everything for a given section (for example, along the vertical axis) and then dropping it to excel OR plotting graphs via plotly or pyplot
Already partially implemented for debug purposes

• Magma should be cooled down by adjacent roof of chamber
• Warmer magma should be cooled down by adjacent colder magma
• Crystals should only form in sufficiently cool magma

• Either via a function color(phase, temperature) that takes a cell and returns an RGB tuple
• Or maybe via Cell class and something like self.getColor

For now they only robustly cover a range from 500 to 2000°C, which is not wide enough