protein-quest 0.5.1__py3-none-any.whl → 0.7.0__py3-none-any.whl

protein_quest/cli.py

@@ -6,14 +6,15 @@ import csv
 import logging
 import os
 import sys
-from collections.abc import Callable, Generator, Iterable
+from collections.abc import Callable, Generator, Iterable, Sequence
+from contextlib import suppress
 from importlib.util import find_spec
-from io import TextIOWrapper
+from io import BytesIO, TextIOWrapper
 from pathlib import Path
 from textwrap import dedent
 
 from cattrs import structure
-from rich import print as rprint
+from rich.console import Console
 from rich.logging import RichHandler
 from rich.markdown import Markdown
 from rich.panel import Panel
@@ -24,18 +25,28 @@ from protein_quest.__version__ import __version__
 from protein_quest.alphafold.confidence import ConfidenceFilterQuery, filter_files_on_confidence
 from protein_quest.alphafold.fetch import DownloadableFormat, downloadable_formats
 from protein_quest.alphafold.fetch import fetch_many as af_fetch
-from protein_quest.converter import converter
+from protein_quest.converter import PositiveInt, converter
 from protein_quest.emdb import fetch as emdb_fetch
 from protein_quest.filters import filter_files_on_chain, filter_files_on_residues
 from protein_quest.go import Aspect, allowed_aspects, search_gene_ontology_term, write_go_terms_to_csv
+from protein_quest.io import (
+    convert_to_cif_files,
+    glob_structure_files,
+    locate_structure_file,
+    read_structure,
+    valid_structure_file_extensions,
+)
 from protein_quest.pdbe import fetch as pdbe_fetch
-from protein_quest.pdbe.io import glob_structure_files, locate_structure_file
 from protein_quest.ss import SecondaryStructureFilterQuery, filter_files_on_secondary_structure
+from protein_quest.structure import structure2uniprot_accessions
 from protein_quest.taxonomy import SearchField, _write_taxonomy_csv, search_fields, search_taxon
 from protein_quest.uniprot import (
     ComplexPortalEntry,
-    PdbResult,
+    PdbResults,
     Query,
+    UniprotDetails,
+    filter_pdb_results_on_chain_length,
+    map_uniprot_accessions2uniprot_details,
     search4af,
     search4emdb,
     search4interaction_partners,
@@ -53,6 +64,8 @@ from protein_quest.utils import (
     user_cache_root_dir,
 )
 
+console = Console(stderr=True)
+rprint = console.print
 logger = logging.getLogger(__name__)
 
 
@@ -93,6 +106,8 @@ def _add_search_uniprot_parser(subparsers: argparse._SubParsersAction):
         action="append",
         help="GO term(s) for molecular function (e.g. GO:0003677). Can be given multiple times.",
     )
+    parser.add_argument("--min-sequence-length", type=int, help="Minimum length of the canonical sequence.")
+    parser.add_argument("--max-sequence-length", type=int, help="Maximum length of the canonical sequence.")
     parser.add_argument("--limit", type=int, default=10_000, help="Maximum number of uniprot accessions to return")
     parser.add_argument("--timeout", type=int, default=1_800, help="Maximum seconds to wait for query to complete")
 
@@ -106,7 +121,7 @@ def _add_search_pdbe_parser(subparsers: argparse._SubParsersAction):
         formatter_class=ArgumentDefaultsRichHelpFormatter,
     )
     parser.add_argument(
-        "uniprot_accs",
+        "uniprot_accessions",
         type=argparse.FileType("r", encoding="UTF-8"),
         help="Text file with UniProt accessions (one per line). Use `-` for stdin.",
     )
@@ -114,15 +129,27 @@ def _add_search_pdbe_parser(subparsers: argparse._SubParsersAction):
         "output_csv",
         type=argparse.FileType("w", encoding="UTF-8"),
         help=dedent("""\
-            Output CSV with `uniprot_acc`, `pdb_id`, `method`, `resolution`, `uniprot_chains`, `chain` columns.
+            Output CSV with following columns:
+            `uniprot_accession`, `pdb_id`, `method`, `resolution`, `uniprot_chains`, `chain`, `chain_length`.
             Where `uniprot_chains` is the raw UniProt chain string, for example `A=1-100`.
-            and where `chain` is the first chain from `uniprot_chains`, for example `A`.
+            and where `chain` is the first chain from `uniprot_chains`, for example `A`
+            and `chain_length` is the length of the chain, for example `100`.
             Use `-` for stdout.
         """),
     )
     parser.add_argument(
         "--limit", type=int, default=10_000, help="Maximum number of PDB uniprot accessions combinations to return"
     )
+    parser.add_argument(
+        "--min-residues",
+        type=int,
+        help="Minimum number of residues required in the chain mapped to the UniProt accession.",
+    )
+    parser.add_argument(
+        "--max-residues",
+        type=int,
+        help="Maximum number of residues allowed in chain mapped to the UniProt accession.",
+    )
     parser.add_argument("--timeout", type=int, default=1_800, help="Maximum seconds to wait for query to complete")
 
 
@@ -135,7 +162,7 @@ def _add_search_alphafold_parser(subparsers: argparse._SubParsersAction):
         formatter_class=ArgumentDefaultsRichHelpFormatter,
     )
     parser.add_argument(
-        "uniprot_accs",
+        "uniprot_accessions",
         type=argparse.FileType("r", encoding="UTF-8"),
         help="Text file with UniProt accessions (one per line). Use `-` for stdin.",
     )
@@ -144,6 +171,8 @@ def _add_search_alphafold_parser(subparsers: argparse._SubParsersAction):
         type=argparse.FileType("w", encoding="UTF-8"),
         help="Output CSV with AlphaFold IDs per UniProt accession. Use `-` for stdout.",
     )
+    parser.add_argument("--min-sequence-length", type=int, help="Minimum length of the canonical sequence.")
+    parser.add_argument("--max-sequence-length", type=int, help="Maximum length of the canonical sequence.")
     parser.add_argument(
         "--limit", type=int, default=10_000, help="Maximum number of Alphafold entry identifiers to return"
     )
@@ -242,7 +271,7 @@ def _add_search_interaction_partners_parser(subparsers: argparse._SubParsersActi
         formatter_class=ArgumentDefaultsRichHelpFormatter,
     )
     parser.add_argument(
-        "uniprot_acc",
+        "uniprot_accession",
         type=str,
         help="UniProt accession (for example P12345).",
     )
@@ -284,7 +313,7 @@ def _add_search_complexes_parser(subparsers: argparse._SubParsersAction):
         formatter_class=ArgumentDefaultsRichHelpFormatter,
     )
     parser.add_argument(
-        "uniprot_accs",
+        "uniprot_accessions",
         type=argparse.FileType("r", encoding="UTF-8"),
         help="Text file with UniProt accessions (one per line) as query for searching complexes. Use `-` for stdin.",
     )
@@ -297,6 +326,76 @@ def _add_search_complexes_parser(subparsers: argparse._SubParsersAction):
     parser.add_argument("--timeout", type=int, default=1_800, help="Maximum seconds to wait for query to complete")
 
 
+def _add_search_uniprot_details_parser(subparsers: argparse._SubParsersAction):
+    """Add search uniprot details subcommand parser."""
+    description = dedent("""\
+        Retrieve UniProt details for given UniProt accessions
+        from the Uniprot SPARQL endpoint.
+
+        The output CSV file has the following columns:
+
+        - uniprot_accession: UniProt accession.
+        - uniprot_id: UniProt ID (mnemonic).
+        - sequence_length: Length of the canonical sequence.
+        - reviewed: Whether the entry is reviewed (Swiss-Prot) or unreviewed (TrEMBL).
+        - protein_name: Recommended protein name.
+        - taxon_id: NCBI Taxonomy ID of the organism.
+        - taxon_name: Scientific name of the organism.
+
+        The order of the output CSV can be different from the input order.
+    """)
+    parser = subparsers.add_parser(
+        "uniprot-details",
+        help="Retrieve UniProt details for given UniProt accessions",
+        description=Markdown(description, style="argparse.text"),  # type: ignore using rich formatter makes this OK
+        formatter_class=ArgumentDefaultsRichHelpFormatter,
+    )
+    parser.add_argument(
+        "uniprot_accessions",
+        type=argparse.FileType("r", encoding="UTF-8"),
+        help="Text file with UniProt accessions (one per line). Use `-` for stdin.",
+    )
+    parser.add_argument(
+        "output_csv",
+        type=argparse.FileType("w", encoding="UTF-8"),
+        help="Output CSV with UniProt details. Use `-` for stdout.",
+    )
+    parser.add_argument("--timeout", type=int, default=1_800, help="Maximum seconds to wait for query to complete")
+    parser.add_argument("--batch-size", type=int, default=1_000, help="Number of accessions to query per batch")
+
+
+def _add_copy_method_arguments(parser):
+    parser.add_argument(
+        "--copy-method",
+        type=str,
+        choices=copy_methods,
+        default="hardlink",
+        help=dedent("""\
+            How to make target file be same file as source file.
+            By default uses hardlinks to save disk space.
+            Note that hardlinks only work within the same filesystem and are harder to track.
+            If you want to track cached files easily then use 'symlink'.
+            On Windows you need developer mode or admin privileges to create symlinks.
+        """),
+    )
+
+
+def _add_cacher_arguments(parser: argparse.ArgumentParser):
+    """Add cacher arguments to parser."""
+    parser.add_argument(
+        "--no-cache",
+        action="store_true",
+        help="Disable caching of files to central location.",
+    )
+    parser.add_argument(
+        "--cache-dir",
+        type=Path,
+        default=user_cache_root_dir(),
+        help="Directory to use as cache for files.",
+    )
+    _add_copy_method_arguments(parser)
+
+
 def _add_retrieve_pdbe_parser(subparsers: argparse._SubParsersAction):
     """Add retrieve pdbe subcommand parser."""
     parser = subparsers.add_parser(
@@ -345,6 +444,19 @@ def _add_retrieve_alphafold_parser(subparsers: argparse._SubParsersAction):
         help=dedent("""AlphaFold formats to retrieve. Can be specified multiple times.
             Default is 'summary' and 'cif'."""),
     )
+    parser.add_argument(
+        "--gzip-files",
+        action="store_true",
+        help="Whether to gzip the downloaded files. Excludes summary files, they are always uncompressed.",
+    )
+    parser.add_argument(
+        "--all-isoforms",
+        action="store_true",
+        help=(
+            "Whether to return all isoforms of each uniprot entry. "
+            "If not given then only the Alphafold entry for the canonical sequence is returned."
+        ),
+    )
     parser.add_argument(
         "--max-parallel-downloads",
         type=int,
@@ -533,6 +645,7 @@ def _add_search_subcommands(subparsers: argparse._SubParsersAction):
     _add_search_taxonomy_parser(subsubparsers)
     _add_search_interaction_partners_parser(subsubparsers)
     _add_search_complexes_parser(subsubparsers)
+    _add_search_uniprot_details_parser(subsubparsers)
 
 
 def _add_retrieve_subcommands(subparsers: argparse._SubParsersAction):
@@ -561,6 +674,75 @@ def _add_filter_subcommands(subparsers: argparse._SubParsersAction):
     _add_filter_ss_parser(subsubparsers)
 
 
+def _add_convert_uniprot_parser(subparsers: argparse._SubParsersAction):
+    """Add convert uniprot subcommand parser."""
+    parser = subparsers.add_parser(
+        "uniprot",
+        help="Convert structure files to list of UniProt accessions.",
+        description="Convert structure files to list of UniProt accessions. "
+        "Uniprot accessions are read from database reference of each structure.",
+        formatter_class=ArgumentDefaultsRichHelpFormatter,
+    )
+    parser.add_argument(
+        "input_dir",
+        type=Path,
+        help=f"Directory with structure files. Supported extensions are {valid_structure_file_extensions}",
+    )
+    parser.add_argument(
+        "output",
+        type=argparse.FileType("wt", encoding="UTF-8"),
+        help="Output text file with UniProt accessions (one per line). Use '-' for stdout.",
+    )
+    parser.add_argument(
+        "--grouped",
+        action="store_true",
+        help="Whether to group accessions by structure file. "
+        "If set output changes to `<structure_file1>,<acc1>\\n<structure_file1>,<acc2>` format.",
+    )
+
+
+def _add_convert_structures_parser(subparsers: argparse._SubParsersAction):
+    """Add convert structures subcommand parser."""
+    parser = subparsers.add_parser(
+        "structures",
+        help="Convert structure files between formats",
+        formatter_class=ArgumentDefaultsRichHelpFormatter,
+    )
+    parser.add_argument(
+        "input_dir",
+        type=Path,
+        help=f"Directory with structure files. Supported extensions are {valid_structure_file_extensions}",
+    )
+    parser.add_argument(
+        "--output-dir",
+        type=Path,
+        help=dedent("""\
+            Directory to write converted structure files. If not given, files are written to `input_dir`.
+        """),
+    )
+    parser.add_argument(
+        "--format",
+        type=str,
+        choices=("cif",),
+        default="cif",
+        help="Output format to convert to.",
+    )
+    _add_copy_method_arguments(parser)
+
+
+def _add_convert_subcommands(subparsers: argparse._SubParsersAction):
+    """Add convert command and its subcommands."""
+    parser = subparsers.add_parser(
+        "convert",
+        help="Convert files between formats",
+        formatter_class=ArgumentDefaultsRichHelpFormatter,
+    )
+    subsubparsers = parser.add_subparsers(dest="convert_cmd", required=True)
+
+    _add_convert_structures_parser(subsubparsers)
+    _add_convert_uniprot_parser(subsubparsers)
+
+
 def _add_mcp_command(subparsers: argparse._SubParsersAction):
     """Add MCP command."""
 
@@ -580,38 +762,6 @@ def _add_mcp_command(subparsers: argparse._SubParsersAction):
     parser.add_argument("--port", default=8000, type=int, help="Port to bind the server to")
 
 
-def _add_copy_method_arguments(parser):
-    parser.add_argument(
-        "--copy-method",
-        type=str,
-        choices=copy_methods,
-        default="hardlink",
-        help=dedent("""\
-            How to make target file be same file as source file.
-            By default uses hardlinks to save disk space.
-            Note that hardlinks only work within the same filesystem and are harder to track.
-            If you want to track cached files easily then use 'symlink'.
-            On Windows you need developer mode or admin privileges to create symlinks.
-        """),
-    )
-
-
-def _add_cacher_arguments(parser: argparse.ArgumentParser):
-    """Add cacher arguments to parser."""
-    parser.add_argument(
-        "--no-cache",
-        action="store_true",
-        help="Disable caching of files to central location.",
-    )
-    parser.add_argument(
-        "--cache-dir",
-        type=Path,
-        default=user_cache_root_dir(),
-        help="Directory to use as cache for files.",
-    )
-    _add_copy_method_arguments(parser)
-
-
 def make_parser() -> argparse.ArgumentParser:
     parser = argparse.ArgumentParser(
         description="Protein Quest CLI", prog="protein-quest", formatter_class=ArgumentDefaultsRichHelpFormatter
@@ -624,25 +774,18 @@ def make_parser() -> argparse.ArgumentParser:
     _add_search_subcommands(subparsers)
     _add_retrieve_subcommands(subparsers)
     _add_filter_subcommands(subparsers)
+    _add_convert_subcommands(subparsers)
     _add_mcp_command(subparsers)
 
     return parser
 
 
-def main():
-    """Main entry point for the CLI."""
-    parser = make_parser()
-    args = parser.parse_args()
-    logging.basicConfig(level=args.log_level, handlers=[RichHandler(show_level=False)])
-
-    # Dispatch table to reduce complexity
-    cmd = args.command
-    sub = getattr(args, f"{cmd}_cmd", None)
-    handler = HANDLERS.get((cmd, sub))
-    if handler is None:
-        msg = f"Unknown command: {cmd} {sub}"
-        raise SystemExit(msg)
-    handler(args)
+def _name_of(file: TextIOWrapper | BytesIO) -> str:
+    try:
+        return file.name
+    except AttributeError:
+        # In pytest BytesIO is used stdout which has no 'name' attribute
+        return "<stdout>"
 
 
 def _handle_search_uniprot(args):
@@ -651,6 +794,8 @@ def _handle_search_uniprot(args):
     subcellular_location_uniprot = args.subcellular_location_uniprot
     subcellular_location_go = args.subcellular_location_go
     molecular_function_go = args.molecular_function_go
+    min_sequence_length = args.min_sequence_length
+    max_sequence_length = args.max_sequence_length
     limit = args.limit
     timeout = args.timeout
     output_file = args.output
@@ -662,54 +807,78 @@ def _handle_search_uniprot(args):
             "subcellular_location_uniprot": subcellular_location_uniprot,
             "subcellular_location_go": subcellular_location_go,
             "molecular_function_go": molecular_function_go,
+            "min_sequence_length": min_sequence_length,
+            "max_sequence_length": max_sequence_length,
         },
         Query,
     )
     rprint("Searching for UniProt accessions")
     accs = search4uniprot(query=query, limit=limit, timeout=timeout)
-    rprint(f"Found {len(accs)} UniProt accessions, written to {output_file.name}")
+    rprint(f"Found {len(accs)} UniProt accessions, written to {_name_of(output_file)}")
     _write_lines(output_file, sorted(accs))
 
 
 def _handle_search_pdbe(args):
-    uniprot_accs = args.uniprot_accs
+    uniprot_accessions = args.uniprot_accessions
     limit = args.limit
     timeout = args.timeout
     output_csv = args.output_csv
+    min_residues = converter.structure(args.min_residues, PositiveInt | None)  # pyright: ignore[reportArgumentType]
+    max_residues = converter.structure(args.max_residues, PositiveInt | None)  # pyright: ignore[reportArgumentType]
 
-    accs = set(_read_lines(uniprot_accs))
+    accs = set(_read_lines(uniprot_accessions))
     rprint(f"Finding PDB entries for {len(accs)} uniprot accessions")
     results = search4pdb(accs, limit=limit, timeout=timeout)
-    total_pdbs = sum([len(v) for v in results.values()])
-    rprint(f"Found {total_pdbs} PDB entries for {len(results)} uniprot accessions")
-    rprint(f"Written to {output_csv.name}")
+
+    raw_nr_results = len(results)
+    raw_total_pdbs = sum([len(v) for v in results.values()])
+    if min_residues or max_residues:
+        results = filter_pdb_results_on_chain_length(results, min_residues, max_residues)
+        total_pdbs = sum([len(v) for v in results.values()])
+        rprint(f"Before filtering found {raw_total_pdbs} PDB entries for {raw_nr_results} uniprot accessions.")
+        rprint(
+            f"After filtering on chain length ({min_residues}, {max_residues}) "
+            f"remained {total_pdbs} PDB entries for {len(results)} uniprot accessions."
+        )
+    else:
+        rprint(f"Found {raw_total_pdbs} PDB entries for {raw_nr_results} uniprot accessions")
+
     _write_pdbe_csv(output_csv, results)
+    rprint(f"Written to {_name_of(output_csv)}")
 
 
 def _handle_search_alphafold(args):
-    uniprot_accs = args.uniprot_accs
+    uniprot_accessions = args.uniprot_accessions
+    min_sequence_length = converter.structure(args.min_sequence_length, PositiveInt | None)  # pyright: ignore[reportArgumentType]
+    max_sequence_length = converter.structure(args.max_sequence_length, PositiveInt | None)  # pyright: ignore[reportArgumentType]
     limit = args.limit
     timeout = args.timeout
     output_csv = args.output_csv
 
-    accs = _read_lines(uniprot_accs)
+    accs = _read_lines(uniprot_accessions)
     rprint(f"Finding AlphaFold entries for {len(accs)} uniprot accessions")
-    results = search4af(accs, limit=limit, timeout=timeout)
-    rprint(f"Found {len(results)} AlphaFold entries, written to {output_csv.name}")
+    results = search4af(
+        accs,
+        min_sequence_length=min_sequence_length,
+        max_sequence_length=max_sequence_length,
+        limit=limit,
+        timeout=timeout,
+    )
+    rprint(f"Found {len(results)} AlphaFold entries, written to {_name_of(output_csv)}")
     _write_dict_of_sets2csv(output_csv, results, "af_id")
 
 
 def _handle_search_emdb(args):
-    uniprot_accs = args.uniprot_accs
+    uniprot_accessions = args.uniprot_accessions
     limit = args.limit
     timeout = args.timeout
     output_csv = args.output_csv
 
-    accs = _read_lines(uniprot_accs)
+    accs = _read_lines(uniprot_accessions)
     rprint(f"Finding EMDB entries for {len(accs)} uniprot accessions")
     results = search4emdb(accs, limit=limit, timeout=timeout)
     total_emdbs = sum([len(v) for v in results.values()])
-    rprint(f"Found {total_emdbs} EMDB entries, written to {output_csv.name}")
+    rprint(f"Found {total_emdbs} EMDB entries, written to {_name_of(output_csv)}")
     _write_dict_of_sets2csv(output_csv, results, "emdb_id")
 
 
@@ -724,7 +893,7 @@ def _handle_search_go(args):
     else:
         rprint(f"Searching for GO terms matching '{term}'")
     results = asyncio.run(search_gene_ontology_term(term, aspect=aspect, limit=limit))
-    rprint(f"Found {len(results)} GO terms, written to {output_csv.name}")
+    rprint(f"Found {len(results)} GO terms, written to {_name_of(output_csv)}")
     write_go_terms_to_csv(results, output_csv)
 
 
@@ -739,36 +908,49 @@ def _handle_search_taxonomy(args):
     else:
         rprint(f"Searching for taxon information matching '{query}'")
     results = asyncio.run(search_taxon(query=query, field=field, limit=limit))
-    rprint(f"Found {len(results)} taxons, written to {output_csv.name}")
+    rprint(f"Found {len(results)} taxons, written to {_name_of(output_csv)}")
     _write_taxonomy_csv(results, output_csv)
 
 
 def _handle_search_interaction_partners(args: argparse.Namespace):
-    uniprot_acc: str = args.uniprot_acc
+    uniprot_accession: str = args.uniprot_accession
     excludes: set[str] = set(args.exclude) if args.exclude else set()
     limit: int = args.limit
     timeout: int = args.timeout
     output_csv: TextIOWrapper = args.output_csv
 
-    rprint(f"Searching for interaction partners of '{uniprot_acc}'")
-    results = search4interaction_partners(uniprot_acc, excludes=excludes, limit=limit, timeout=timeout)
-    rprint(f"Found {len(results)} interaction partners, written to {output_csv.name}")
+    rprint(f"Searching for interaction partners of '{uniprot_accession}'")
+    results = search4interaction_partners(uniprot_accession, excludes=excludes, limit=limit, timeout=timeout)
+    rprint(f"Found {len(results)} interaction partners, written to {_name_of(output_csv)}")
     _write_lines(output_csv, results.keys())
 
 
 def _handle_search_complexes(args: argparse.Namespace):
-    uniprot_accs = args.uniprot_accs
+    uniprot_accessions = args.uniprot_accessions
     limit = args.limit
     timeout = args.timeout
     output_csv = args.output_csv
 
-    accs = _read_lines(uniprot_accs)
+    accs = _read_lines(uniprot_accessions)
     rprint(f"Finding complexes for {len(accs)} uniprot accessions")
     results = search4macromolecular_complexes(accs, limit=limit, timeout=timeout)
-    rprint(f"Found {len(results)} complexes, written to {output_csv.name}")
+    rprint(f"Found {len(results)} complexes, written to {_name_of(output_csv)}")
     _write_complexes_csv(results, output_csv)
 
 
+def _handle_search_uniprot_details(args: argparse.Namespace):
+    uniprot_accessions = args.uniprot_accessions
+    timeout = args.timeout
+    batch_size = args.batch_size
+    output_csv: TextIOWrapper = args.output_csv
+
+    accs = _read_lines(uniprot_accessions)
+    rprint(f"Retrieving UniProt entry details for {len(accs)} uniprot accessions")
+    results = list(map_uniprot_accessions2uniprot_details(accs, timeout=timeout, batch_size=batch_size))
+    _write_uniprot_details_csv(output_csv, results)
+    rprint(f"Retrieved details for {len(results)} UniProt entries, written to {_name_of(output_csv)}")
+
+
 def _initialize_cacher(args: argparse.Namespace) -> Cacher:
     if args.no_cache:
         return PassthroughCacher()
@@ -798,17 +980,25 @@ def _handle_retrieve_alphafold(args):
     alphafold_csv = args.alphafold_csv
     max_parallel_downloads = args.max_parallel_downloads
     cacher = _initialize_cacher(args)
+    gzip_files = args.gzip_files
+    all_isoforms = args.all_isoforms
 
     if what_formats is None:
         what_formats = {"summary", "cif"}
 
-    # TODO besides `uniprot_acc,af_id\n` csv also allow headless single column format
+    # TODO besides `uniprot_accession,af_id\n` csv also allow headless single column format
     #
     af_ids = _read_column_from_csv(alphafold_csv, "af_id")
     validated_what: set[DownloadableFormat] = structure(what_formats, set[DownloadableFormat])
     rprint(f"Retrieving {len(af_ids)} AlphaFold entries with formats {validated_what}")
     afs = af_fetch(
-        af_ids, download_dir, what=validated_what, max_parallel_downloads=max_parallel_downloads, cacher=cacher
+        af_ids,
+        download_dir,
+        what=validated_what,
+        max_parallel_downloads=max_parallel_downloads,
+        cacher=cacher,
+        gzip_files=gzip_files,
+        all_isoforms=all_isoforms,
     )
     total_nr_files = sum(af.nr_of_files() for af in afs)
     rprint(f"Retrieved {total_nr_files} AlphaFold files and {len(afs)} summaries, written to {download_dir}")
@@ -863,11 +1053,11 @@ def _handle_filter_confidence(args: argparse.Namespace):
         if r.filtered_file:
             passed_count += 1
         if stats_file:
-            writer.writerow([r.input_file, r.count, r.filtered_file is not None, r.filtered_file])
+            writer.writerow([r.input_file, r.count, r.filtered_file is not None, r.filtered_file])  # pyright: ignore[reportPossiblyUnboundVariable]
 
     rprint(f"Filtered {passed_count} mmcif/PDB files by confidence, written to {output_dir} directory")
     if stats_file:
-        rprint(f"Statistics written to {stats_file.name}")
+        rprint(f"Statistics written to {_name_of(stats_file)}")
 
 
 def _handle_filter_chain(args):
@@ -933,13 +1123,13 @@ def _handle_filter_residue(args):
         input_files, output_dir, min_residues=min_residues, max_residues=max_residues, copy_method=copy_method
     ):
         if stats_file:
-            writer.writerow([r.input_file, r.residue_count, r.passed, r.output_file])
+            writer.writerow([r.input_file, r.residue_count, r.passed, r.output_file])  # pyright: ignore[reportPossiblyUnboundVariable]
         if r.passed:
             nr_passed += 1
 
     rprint(f"Wrote {nr_passed} files to {output_dir} directory.")
     if stats_file:
-        rprint(f"Statistics written to {stats_file.name}")
+        rprint(f"Statistics written to {_name_of(stats_file)}")
 
 
 def _handle_filter_ss(args):
@@ -987,7 +1177,7 @@ def _handle_filter_ss(args):
             copyfile(input_file, output_file, copy_method)
             nr_passed += 1
         if stats_file:
-            writer.writerow(
+            writer.writerow(  # pyright: ignore[reportPossiblyUnboundVariable]
                 [
                     input_file,
                     result.stats.nr_residues,
@@ -1001,7 +1191,7 @@ def _handle_filter_ss(args):
             )
     rprint(f"Wrote {nr_passed} files to {output_dir} directory.")
     if stats_file:
-        rprint(f"Statistics written to {stats_file.name}")
+        rprint(f"Statistics written to {_name_of(stats_file)}")
 
 
 def _handle_mcp(args):
@@ -1017,24 +1207,45 @@ def _handle_mcp(args):
         mcp.run(transport=args.transport, host=args.host, port=args.port)
 
 
-HANDLERS: dict[tuple[str, str | None], Callable] = {
-    ("search", "uniprot"): _handle_search_uniprot,
-    ("search", "pdbe"): _handle_search_pdbe,
-    ("search", "alphafold"): _handle_search_alphafold,
-    ("search", "emdb"): _handle_search_emdb,
-    ("search", "go"): _handle_search_go,
-    ("search", "taxonomy"): _handle_search_taxonomy,
-    ("search", "interaction-partners"): _handle_search_interaction_partners,
-    ("search", "complexes"): _handle_search_complexes,
-    ("retrieve", "pdbe"): _handle_retrieve_pdbe,
-    ("retrieve", "alphafold"): _handle_retrieve_alphafold,
-    ("retrieve", "emdb"): _handle_retrieve_emdb,
-    ("filter", "confidence"): _handle_filter_confidence,
-    ("filter", "chain"): _handle_filter_chain,
-    ("filter", "residue"): _handle_filter_residue,
-    ("filter", "secondary-structure"): _handle_filter_ss,
-    ("mcp", None): _handle_mcp,
-}
+def _handle_convert_uniprot(args):
+    input_dir = structure(args.input_dir, Path)
+    output_file: TextIOWrapper = args.output
+    grouped: bool = args.grouped
+    input_files = sorted(glob_structure_files(input_dir))
+    if grouped:
+        for input_file in tqdm(input_files, unit="file"):
+            s = read_structure(input_file)
+            uniprot_accessions = structure2uniprot_accessions(s)
+            _write_lines(
+                output_file, [f"{input_file},{uniprot_accession}" for uniprot_accession in sorted(uniprot_accessions)]
+            )
+    else:
+        uniprot_accessions: set[str] = set()
+        for input_file in tqdm(input_files, unit="file"):
+            s = read_structure(input_file)
+            uniprot_accessions.update(structure2uniprot_accessions(s))
+        _write_lines(output_file, sorted(uniprot_accessions))
+
+
+def _handle_convert_structures(args):
+    input_dir = structure(args.input_dir, Path)
+    output_dir = input_dir if args.output_dir is None else structure(args.output_dir, Path)
+    output_dir.mkdir(parents=True, exist_ok=True)
+    copy_method: CopyMethod = structure(args.copy_method, CopyMethod)  # pyright: ignore[reportArgumentType]
+
+    input_files = sorted(glob_structure_files(input_dir))
+    rprint(f"Converting {len(input_files)} files in {input_dir} directory to cif format.")
+    for _ in tqdm(
+        convert_to_cif_files(
+            input_files,
+            output_dir,
+            copy_method=copy_method,
+        ),
+        total=len(input_files),
+        unit="file",
+    ):
+        pass
+    rprint(f"Converted {len(input_files)} files into {output_dir}.")
 
 
 def _read_lines(file: TextIOWrapper) -> list[str]:
@@ -1042,7 +1253,8 @@ def _read_lines(file: TextIOWrapper) -> list[str]:
 
 
 def _make_sure_parent_exists(file: TextIOWrapper):
-    if file.name != "<stdout>":
+    # Can not create dir for stdout
+    with suppress(AttributeError):
         Path(file.name).parent.mkdir(parents=True, exist_ok=True)
 
 
@@ -1051,34 +1263,35 @@ def _write_lines(file: TextIOWrapper, lines: Iterable[str]):
     file.writelines(line + os.linesep for line in lines)
 
 
-def _write_pdbe_csv(path: TextIOWrapper, data: dict[str, set[PdbResult]]):
+def _write_pdbe_csv(path: TextIOWrapper, data: PdbResults):
     _make_sure_parent_exists(path)
-    fieldnames = ["uniprot_acc", "pdb_id", "method", "resolution", "uniprot_chains", "chain"]
+    fieldnames = ["uniprot_accession", "pdb_id", "method", "resolution", "uniprot_chains", "chain", "chain_length"]
     writer = csv.DictWriter(path, fieldnames=fieldnames)
     writer.writeheader()
-    for uniprot_acc, entries in sorted(data.items()):
+    for uniprot_accession, entries in sorted(data.items()):
         for e in sorted(entries, key=lambda x: (x.id, x.method)):
             writer.writerow(
                 {
-                    "uniprot_acc": uniprot_acc,
+                    "uniprot_accession": uniprot_accession,
                     "pdb_id": e.id,
                     "method": e.method,
                     "resolution": e.resolution or "",
                     "uniprot_chains": e.uniprot_chains,
                     "chain": e.chain,
+                    "chain_length": e.chain_length,
                 }
             )
 
 
 def _write_dict_of_sets2csv(file: TextIOWrapper, data: dict[str, set[str]], ref_id_field: str):
     _make_sure_parent_exists(file)
-    fieldnames = ["uniprot_acc", ref_id_field]
+    fieldnames = ["uniprot_accession", ref_id_field]
 
     writer = csv.DictWriter(file, fieldnames=fieldnames)
     writer.writeheader()
-    for uniprot_acc, ref_ids in sorted(data.items()):
+    for uniprot_accession, ref_ids in sorted(data.items()):
         for ref_id in sorted(ref_ids):
-            writer.writerow({"uniprot_acc": uniprot_acc, ref_id_field: ref_id})
+            writer.writerow({"uniprot_accession": uniprot_accession, ref_id_field: ref_id})
 
 
 def _iter_csv_rows(file: TextIOWrapper) -> Generator[dict[str, str]]:
@@ -1118,3 +1331,61 @@ def _write_complexes_csv(complexes: list[ComplexPortalEntry], output_csv: TextIO
                 members_str,
             ]
         )
+
+
+def _write_uniprot_details_csv(
+    output_csv: TextIOWrapper,
+    uniprot_details_list: Iterable[UniprotDetails],
+) -> None:
+    if not uniprot_details_list:
+        msg = "No UniProt entries found for given accessions"
+        raise ValueError(msg)
+    # As all props of UniprotDetails are scalar, we can directly unstructure to dicts
+    rows = converter.unstructure(uniprot_details_list)
+    fieldnames = rows[0].keys()
+    writer = csv.DictWriter(output_csv, fieldnames=fieldnames)
+    writer.writeheader()
+    writer.writerows(rows)
+
+
+HANDLERS: dict[tuple[str, str | None], Callable] = {
+    ("search", "uniprot"): _handle_search_uniprot,
+    ("search", "pdbe"): _handle_search_pdbe,
+    ("search", "alphafold"): _handle_search_alphafold,
+    ("search", "emdb"): _handle_search_emdb,
+    ("search", "go"): _handle_search_go,
+    ("search", "taxonomy"): _handle_search_taxonomy,
+    ("search", "interaction-partners"): _handle_search_interaction_partners,
+    ("search", "complexes"): _handle_search_complexes,
+    ("search", "uniprot-details"): _handle_search_uniprot_details,
+    ("retrieve", "pdbe"): _handle_retrieve_pdbe,
+    ("retrieve", "alphafold"): _handle_retrieve_alphafold,
+    ("retrieve", "emdb"): _handle_retrieve_emdb,
+    ("filter", "confidence"): _handle_filter_confidence,
+    ("filter", "chain"): _handle_filter_chain,
+    ("filter", "residue"): _handle_filter_residue,
+    ("filter", "secondary-structure"): _handle_filter_ss,
+    ("mcp", None): _handle_mcp,
+    ("convert", "structures"): _handle_convert_structures,
+    ("convert", "uniprot"): _handle_convert_uniprot,
+}
+
+
+def main(argv: Sequence[str] | None = None):
+    """Main entry point for the CLI.
+
+    Args:
+        argv: List of command line arguments. If None, uses sys.argv.
+    """
+    parser = make_parser()
+    args = parser.parse_args(argv)
+    logging.basicConfig(level=args.log_level, handlers=[RichHandler(show_level=False, console=console)])
+
+    # Dispatch table to reduce complexity
+    cmd = args.command
+    sub = getattr(args, f"{cmd}_cmd", None)
+    handler = HANDLERS.get((cmd, sub))
+    if handler is None:
+        msg = f"Unknown command: {cmd} {sub}"
+        raise SystemExit(msg)
+    handler(args)