chore: use dedicated types for compressed modswitched conformance

[nsarlin-zama]

PR content/description

modswitched expansion type info was stored into the CiphertextConformanceParams, which meant that we had to provide this information even for ciphertexts that were not compressed. This PR adds dedicated types for modswitched ciphertexts conformance params.

---

This change is 

[mayeul-zama]

@mayeul-zama reviewed 20 of 20 files at r1, all commit messages.
Reviewable status: all files reviewed, 5 unresolved discussions (waiting on @IceTDrinker and @tmontaigu)

---

tfhe/src/core_crypto/entities/compressed_modulus_switched_multi_bit_lwe_ciphertext.rs line 541 at r1 (raw file):

                .as_ref()
                .is_none_or(|packed_diffs| packed_diffs.is_conformant(&lwe_dim))
            && *lwe_dimension == ct_params.lwe_dim

Could also destructure ct_params

---

tfhe/src/core_crypto/entities/compressed_modulus_switched_lwe_ciphertext.rs line 150 at r1 (raw file):

{
    pub ct_params: LweCiphertextConformanceParams<Scalar>,
    pub ms_decompression_method: MsDecompressionType,

"Type" vs "method", would be clearer to use only one name
I have no personal preference

---

tfhe/src/core_crypto/entities/compressed_modulus_switched_lwe_ciphertext.rs line 168 at r1 (raw file):

        let CompressedModulusSwitchedLweCiphertextConformanceParams {
            ct_params,

Could also destructure ct_params

---

tfhe/src/high_level_api/integers/unsigned/compressed.rs line 226 at r1 (raw file):

);

impl CompressedRadixCiphertextConformanceParams {}

To remove

---

tfhe/src/shortint/ciphertext/compressed_modulus_switched_ciphertext.rs line 86 at r1 (raw file):

        } = self;

        compressed_modulus_switched_lwe_ciphertext.is_conformant(&param.ct_params)

Not really releveant to this PR

param could be destructured

I don't know how time consuming it would be to make sure we follow this pattern in all conformance code, maybe not so much

[nsarlin-zama]

Reviewable status: 11 of 23 files reviewed, 5 unresolved discussions (waiting on @IceTDrinker, @mayeul-zama, and @tmontaigu)

---

tfhe/src/core_crypto/entities/compressed_modulus_switched_lwe_ciphertext.rs line 150 at r1 (raw file):

Previously, mayeul-zama wrote…

"Type" vs "method", would be clearer to use only one name
I have no personal preference

done

---

tfhe/src/core_crypto/entities/compressed_modulus_switched_lwe_ciphertext.rs line 168 at r1 (raw file):

Previously, mayeul-zama wrote…

Could also destructure ct_params

done

---

tfhe/src/core_crypto/entities/compressed_modulus_switched_multi_bit_lwe_ciphertext.rs line 541 at r1 (raw file):

Previously, mayeul-zama wrote…

Could also destructure ct_params

done

---

tfhe/src/high_level_api/integers/unsigned/compressed.rs line 226 at r1 (raw file):

Previously, mayeul-zama wrote…

To remove

done

---

tfhe/src/shortint/ciphertext/compressed_modulus_switched_ciphertext.rs line 86 at r1 (raw file):

Previously, mayeul-zama wrote…

Not really releveant to this PR

param could be destructured

I don't know how time consuming it would be to make sure we follow this pattern in all conformance code, maybe not so much

done, required a small refacto to remove an unused field

[mayeul-zama]

Thanks!

@mayeul-zama reviewed 4 of 4 files at r2, 8 of 8 files at r3, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @IceTDrinker and @tmontaigu)

[mayeul-zama]

@mayeul-zama reviewed 2 of 2 files at r4, all commit messages.
Reviewable status: complete! all files reviewed, all discussions resolved (waiting on @IceTDrinker and @tmontaigu)