File size: 8,618 Bytes
44459bb |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 |
"""Query module for Chai prediction endpoint."""
from __future__ import annotations
import shutil
import tempfile
from io import StringIO
from itertools import chain
from pathlib import Path
from typing import Any
from folding_studio_data_models import FoldingModel
from pydantic import BaseModel, field_validator
from folding_studio.commands.utils import (
a3m_to_aligned_pqt,
process_uploaded_msas,
)
from folding_studio.query import Query
from folding_studio.utils.fasta import validate_fasta
from folding_studio.utils.headers import get_auth_headers
from folding_studio.utils.path_helpers import validate_path
class ChaiParameters(BaseModel):
"""Chai1 inference parameters."""
seed: int = 0
num_trunk_recycles: int = 3
num_diffn_timesteps: int = 200
recycle_msa_subsample: int = 0
num_trunk_samples: int = 1
restraints: str | None = None
use_msa_server: bool = False
use_templates_server: bool = False
custom_msa_paths: dict[str, str] | None = None
@field_validator("restraints", mode="before")
def read_restraints(
cls: ChaiParameters, restraints: str | Path | None
) -> str | None:
"""Reads restraints from a CSV file and returns its content as a string."""
if restraints is None:
return
path = validate_path(restraints, is_file=True, file_suffix=(".csv"))
with path.open(newline="", encoding="utf-8") as csvfile:
return csvfile.read().strip()
class ChaiQuery(Query):
"""Chai1 model query."""
MODEL = FoldingModel.CHAI
def __init__(
self,
fasta_files: dict[str, str],
query_name: str,
parameters: ChaiParameters = ChaiParameters(),
):
"""Initializes a ChaiQuery instance."""
if not fasta_files:
raise ValueError("FASTA files dictionary cannot be empty.")
self.fasta_files = fasta_files
self.query_name = query_name
self._parameters = parameters
@classmethod
def from_protein_sequence(
cls: ChaiQuery, sequence: str, query_name: str | None = None, **kwargs
) -> ChaiQuery:
"""Initialize a ChaiQuery instance from a str protein sequence.
Args:
sequence (str): Protein amino-acid sequence
query_name (str | None, optional): User-defined query name. Defaults to None.
Raises:
NotAMonomer: If the sequence is not a monomer complex.
Returns:
ChaiQuery
"""
record = validate_fasta(StringIO(sequence))
custom_msa_paths = kwargs.pop("custom_msa_paths", None)
if custom_msa_paths:
kwargs["custom_msa_paths"] = cls._upload_custom_msa_files(custom_msa_paths)
query_name = (
query_name
if query_name is not None
else record.description.split("|", maxsplit=1)[0] # first tag
)
return cls(
fasta_files={query_name: sequence},
query_name=query_name,
parameters=ChaiParameters(**kwargs),
)
@classmethod
def from_file(
cls: ChaiQuery, path: str | Path, query_name: str | None = None, **kwargs
) -> ChaiQuery:
"""Initialize a ChaiQuery instance from a file.
Supported file format are:
- FASTA
Args:
path (str | Path): Path of the FASTA file.
query_name (str | None, optional): User-defined query name. Defaults to None.
Returns:
ChaiQuery
"""
path = validate_path(path, is_file=True, file_suffix=(".fasta", ".fa"))
custom_msa_paths = kwargs.pop("custom_msa_paths", None)
if custom_msa_paths:
kwargs["custom_msa_paths"] = cls._upload_custom_msa_files(custom_msa_paths)
query_name = query_name or path.stem
return cls(
fasta_files={path.stem: validate_fasta(path, str_output=True)},
query_name=query_name,
parameters=ChaiParameters(**kwargs),
)
@classmethod
def from_directory(
cls: ChaiQuery, path: str | Path, query_name: str | None = None, **kwargs
) -> ChaiQuery:
"""Initialize a ChaiQuery instance from a directory.
Supported file format in directory are:
- FASTA
Args:
path (str | Path): Path to a directory of FASTA files.
query_name (str | None, optional): User-defined query name. Defaults to None.
seed (int, optional): Random seed. Defaults to 0.
Raises:
ValueError: If no FASTA file are present in the directory.
Returns:
ChaiQuery
"""
path = validate_path(path, is_dir=True)
custom_msa_paths = kwargs.pop("custom_msa_paths", None)
if custom_msa_paths:
kwargs["custom_msa_paths"] = cls._upload_custom_msa_files(custom_msa_paths)
print(kwargs["custom_msa_paths"])
fasta_files = {
file.stem: validate_fasta(file, str_output=True)
for file in chain(path.glob("*.fasta"), path.glob("*.fa"))
}
if not fasta_files:
raise ValueError(f"No FASTA files found in directory '{path}'.")
query_name = query_name or path.name
return cls(
fasta_files=fasta_files,
query_name=query_name,
parameters=ChaiParameters(**kwargs),
)
@property
def payload(self) -> dict[str, Any]:
"""Payload to send to the prediction API endpoint."""
return {
"fasta_files": self.fasta_files,
"use_msa_server": self.parameters.use_msa_server,
"use_templates_server": self.parameters.use_templates_server,
"num_trunk_recycles": self.parameters.num_trunk_recycles,
"seed": self.parameters.seed,
"num_diffn_timesteps": self.parameters.num_diffn_timesteps,
"restraints": self.parameters.restraints,
"recycle_msa_subsample": self.parameters.recycle_msa_subsample,
"num_trunk_samples": self.parameters.num_trunk_samples,
"custom_msa_paths": self.parameters.custom_msa_paths,
}
@property
def parameters(self) -> ChaiParameters:
"""Parameters of the query."""
return self._parameters
@staticmethod
def _upload_custom_msa_files(
source: str, headers: str | None = None
) -> dict[str, str]:
"""Read A3M or MSA files from a file or directory and uploads them to GCS.
Args:
source (str): Path to an .a3m or .aligned.pqt file or a directory containing .a3m or .aligned.pqt files
identity_token (str | None, optional): GCP identity token. Defaults to None.
Raises:
ValueError: If file has unsupported extension.
ValueError: If directory has no supported file.
Returns:
dict[str, str]: _description_
"""
headers = headers or get_auth_headers()
source_path = validate_path(source)
# Process if source is a file.
if source_path.is_file():
if source_path.suffix == ".a3m":
with tempfile.TemporaryDirectory() as tmpdir:
tmp_path = Path(tmpdir)
shutil.copy(source_path, tmp_path / source_path.name)
pqt_file = a3m_to_aligned_pqt(str(tmp_path))
return process_uploaded_msas([Path(pqt_file)], headers)
elif source_path.name.endswith(".aligned.pqt"):
return process_uploaded_msas([source_path], headers)
else:
raise ValueError(
f"Invalid file type: {source_path.suffix}. Expected '.a3m' or a file ending with '.aligned.pqt'."
)
# Process if source is a directory.
elif source_path.is_dir():
pqt_files = list(source_path.glob("*.aligned.pqt"))
if pqt_files:
return process_uploaded_msas(pqt_files, headers)
a3m_files = list(source_path.glob("*.a3m"))
if not a3m_files:
raise ValueError(
f"Directory '{source}' contains no files ending with '.aligned.pqt' or '.a3m'."
)
with tempfile.TemporaryDirectory() as tmpdir:
tmp_path = Path(tmpdir)
for file in a3m_files:
shutil.copy(file, tmp_path / file.name)
pqt_file = a3m_to_aligned_pqt(str(tmp_path))
return process_uploaded_msas([Path(pqt_file)], headers)
|