feat: solve strategies

[baszalmstra]

Implements solver strategies:

• Highest (default): Solve all packages to the highest compatible version.
• LowestVersion: Solve all packages to the lowest compatible version. This only looks at the version. This ensures that build number ordering and downprioritization remain functional.
• LowestVersionDirect: Solve for the lowest compatible version, similar to LowestVersion, but only for direct dependencies. Transitive dependencies are still sorted using the Highest strategy.

This is only implemented when using resolvo. When using a non-default strategy with libsolv you will get an Unsupported error.

This is also implemented into the binary:

> cargo run -- create --dry-run 'python>2' --platform linux-64 --strategy highest
+ _libgcc_mutex 0.1 conda_forge
+ libgomp 13.2.0 h77fa898_7
+ _openmp_mutex 4.5 2_gnu
+ libgcc-ng 13.2.0 h77fa898_7
+ xz 5.2.6 h166bdaf_0
+ tzdata 2024a h0c530f3_0
+ libzlib 1.2.13 hd590300_5
+ tk 8.6.13 noxft_h4845f30_101
+ ncurses 6.5 h59595ed_0
+ readline 8.2 h8228510_1
+ ca-certificates 2024.2.2 hbcca054_0
+ openssl 3.3.0 hd590300_0
+ libxcrypt 4.4.36 hd590300_1
+ libuuid 2.38.1 h0b41bf4_0
+ libsqlite 3.45.3 h2797004_0
+ libnsl 2.0.1 hd590300_0
+ libffi 3.4.2 h7f98852_5
+ libexpat 2.6.2 h59595ed_0
+ ld_impl_linux-64 2.40 h55db66e_0
+ bzip2 1.0.8 hd590300_5
+ python 3.12.3 hab00c5b_0_cpython

> cargo run -- create --dry-run 'python>2' --platform linux-64 --strategy lowest
+ zlib 1.2.8 3
+ tk 8.5.19 2
+ sqlite 3.13.0 1
+ ncurses 5.9 10
+ readline 6.2 0
+ ca-certificates 2016.2.28 0
+ openssl 1.0.2h 3
+ python 2.6.9 0

> cargo run -- create --dry-run 'python>2' --platform linux-64 --strategy lowest-direct
+ _libgcc_mutex 0.1 conda_forge
+ libgomp 13.2.0 h77fa898_7
+ _openmp_mutex 4.5 2_gnu
+ libgcc-ng 13.2.0 h77fa898_7
+ libzlib 1.2.13 hd590300_5
+ zlib 1.2.13 hd590300_5
+ tk 8.5.19 2
+ sqlite 3.13.0 1
+ ncurses 5.9 10
+ readline 6.2 0
+ ca-certificates 2024.2.2 hbcca054_0
+ openssl 1.0.2u h516909a_0
+ python 2.6.9 0

Note that Im using python>2 because there is also a python 1.0.1 on conda-forge.

Fixes #330

[wolfv]

Looks great!