protein-quest 0.3.0__py3-none-any.whl → 0.3.2__py3-none-any.whl

protein_quest/utils.py

@@ -2,20 +2,24 @@
 
 import asyncio
 import logging
-from collections.abc import Iterable
+import shutil
+from collections.abc import Coroutine, Iterable
 from contextlib import asynccontextmanager
 from pathlib import Path
+from textwrap import dedent
+from typing import Any, Literal, get_args
 
 import aiofiles
 import aiohttp
 from aiohttp_retry import ExponentialRetry, RetryClient
 from tqdm.asyncio import tqdm
+from yarl import URL
 
 logger = logging.getLogger(__name__)
 
 
 async def retrieve_files(
-    urls: Iterable[tuple[str, str]],
+    urls: Iterable[tuple[URL | str, str]],
     save_dir: Path,
     max_parallel_downloads: int = 5,
     retries: int = 3,
@@ -45,7 +49,7 @@ async def retrieve_files(
 
 async def _retrieve_file(
     session: RetryClient,
-    url: str,
+    url: URL | str,
     save_path: Path,
     semaphore: asyncio.Semaphore,
     ovewrite: bool = False,
@@ -103,3 +107,61 @@ async def friendly_session(retries: int = 3, total_timeout: int = 300):
     async with aiohttp.ClientSession(timeout=timeout) as session:
         client = RetryClient(client_session=session, retry_options=retry_options)
         yield client
+
+
+class NestedAsyncIOLoopError(RuntimeError):
+    """Custom error for nested async I/O loops."""
+
+    def __init__(self) -> None:
+        msg = dedent("""\
+            Can not run async method from an environment where the asyncio event loop is already running.
+            Like a Jupyter notebook.
+
+            Please use the async function directly or
+            call `import nest_asyncio; nest_asyncio.apply()` and try again.
+            """)
+        super().__init__(msg)
+
+
+def run_async[R](coroutine: Coroutine[Any, Any, R]) -> R:
+    """Run an async coroutine with nicer error.
+
+    Args:
+        coroutine: The async coroutine to run.
+
+    Returns:
+        The result of the coroutine.
+
+    Raises:
+        NestedAsyncIOLoopError: If called from a nested async I/O loop like in a Jupyter notebook.
+    """
+    try:
+        return asyncio.run(coroutine)
+    except RuntimeError as e:
+        raise NestedAsyncIOLoopError from e
+
+
+CopyMethod = Literal["copy", "symlink"]
+copy_methods = set(get_args(CopyMethod))
+
+
+def copyfile(source: Path, target: Path, copy_method: CopyMethod = "copy"):
+    """Make target path be same file as source by either copying or symlinking.
+
+    Args:
+        source: The source file to copy or symlink.
+        target: The target file to create.
+        copy_method: The method to use for copying.
+
+    Raises:
+        FileNotFoundError: If the source file or parent of target does not exist.
+        ValueError: If the method is not "copy" or "symlink".
+    """
+    if copy_method == "copy":
+        shutil.copyfile(source, target)
+    elif copy_method == "symlink":
+        rel_source = source.relative_to(target.parent, walk_up=True)
+        target.symlink_to(rel_source)
+    else:
+        msg = f"Unknown method: {copy_method}"
+        raise ValueError(msg)

{protein_quest-0.3.0.dist-info → protein_quest-0.3.2.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: protein_quest
-Version: 0.3.0
+Version: 0.3.2
 Summary: Search/retrieve/filter proteins and protein structures
 Project-URL: Homepage, https://github.com/haddocking/protein-quest
 Project-URL: Issues, https://github.com/haddocking/protein-quest/issues
@@ -13,19 +13,16 @@ Requires-Dist: aiohttp-retry>=2.9.1
 Requires-Dist: aiohttp[speedups]>=3.11.18
 Requires-Dist: aiopath>=0.7.7
 Requires-Dist: attrs>=25.3.0
-Requires-Dist: bokeh>=3.7.3
 Requires-Dist: cattrs[orjson]>=24.1.3
 Requires-Dist: dask>=2025.5.1
 Requires-Dist: distributed>=2025.5.1
 Requires-Dist: gemmi>=0.7.3
-Requires-Dist: molviewspec>=1.6.0
-Requires-Dist: pandas>=2.3.0
-Requires-Dist: platformdirs>=4.3.8
 Requires-Dist: psutil>=7.0.0
 Requires-Dist: rich-argparse>=1.7.1
 Requires-Dist: rich>=14.0.0
 Requires-Dist: sparqlwrapper>=2.0.0
 Requires-Dist: tqdm>=4.67.1
+Requires-Dist: yarl>=1.20.1
 Provides-Extra: mcp
 Requires-Dist: fastmcp>=2.11.3; extra == 'mcp'
 Requires-Dist: pydantic>=2.11.7; extra == 'mcp'
@@ -37,8 +34,7 @@ Description-Content-Type: text/markdown
 [![CI](https://github.com/haddocking/protein-quest/actions/workflows/ci.yml/badge.svg)](https://github.com/haddocking/protein-quest/actions/workflows/ci.yml)
 [![Research Software Directory Badge](https://img.shields.io/badge/rsd-00a3e3.svg)](https://www.research-software.nl/software/protein-quest)
 [![PyPI](https://img.shields.io/pypi/v/protein-quest)](https://pypi.org/project/protein-quest/)
-<!-- TODO replace with correct zenodo id -->
-[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.15632658.svg)](https://doi.org/10.5281/zenodo.15632658)
+[![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.16941288.svg)](https://doi.org/10.5281/zenodo.16941288)
 [![Codacy Badge](https://app.codacy.com/project/badge/Coverage/7a3f3f1fe64640d583a5e50fe7ba828e)](https://app.codacy.com/gh/haddocking/protein-quest/coverage?utm_source=gh&utm_medium=referral&utm_content=&utm_campaign=Badge_coverage)
 
 Python package to search/retrieve/filter proteins and protein structures.
@@ -63,9 +59,11 @@ graph TB;
     searchpdbe -->|pdb_ids|fetchpdbe[Retrieve PDBe]
     searchaf --> |uniprot_accessions|fetchad(Retrieve AlphaFold)
     searchemdb -. emdb_ids .->fetchemdb[Retrieve EMDB]
-    fetchpdbe -->|mmcif_files_with_uniprot_acc| chainfilter{Filter on chain of uniprot}
-    chainfilter --> |mmcif_files| residuefilter{Filter on chain length}
-    fetchad -->|pdb_files| confidencefilter{Filter out low confidence}
+    fetchpdbe -->|mmcif_files_with_uniprot_acc| chainfilter{{Filter on chain of uniprot}}
+    chainfilter --> |mmcif_files| residuefilter{{Filter on chain length}}
+    fetchad -->|pdb_files| confidencefilter{{Filter out low confidence}}
+    confidencefilter --> |mmcif_files| ssfilter{{Filter on secondary structure}}
+    residuefilter --> |mmcif_files| ssfilter
     classDef dashedBorder stroke-dasharray: 5 5;
     goterm:::dashedBorder
     taxonomy:::dashedBorder
@@ -90,7 +88,7 @@ pip install git+https://github.com/haddocking/protein-quest.git
 
 The main entry point is the `protein-quest` command line tool which has multiple subcommands to perform actions.
 
-To use programmaticly, see [API documentation](https://www.bonvinlab.org/protein-quest/autoapi/summary/).
+To use programmaticly, see the [Jupyter notebooks](https://www.bonvinlab.org/protein-quest/notebooks) and [API documentation](https://www.bonvinlab.org/protein-quest/autoapi/summary/).
 
 ### Search Uniprot accessions
 
@@ -179,6 +177,18 @@ protein-quest filter residue  \
     ./filtered-chains ./filtered
 ```
 
+### To filter on secondary structure
+
+To filter on structure being mostly alpha helices and have no beta sheets.
+
+```shell
+protein-quest filter secondary-structure \
+    --ratio-min-helix-residues 0.5 \
+    --ratio-max-sheet-residues 0.0 \
+    --write-stats filtered-ss/stats.csv \
+    ./filtered-chains ./filtered-ss
+```
+
 ### Search Taxonomy
 
 ```shell

protein_quest-0.3.2.dist-info/RECORD

@@ -0,0 +1,26 @@
+protein_quest/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+protein_quest/__version__.py,sha256=tDIN8WjNdFKRoXsf6tArV0_n6nbcPEBWNv1zuhaRbKo,56
+protein_quest/cli.py,sha256=k4HC282QkbAAIk614vIJgaKfkS3XD9hYj7E5hEuiDxA,37893
+protein_quest/converter.py,sha256=tSDw7HOlC7UoWryr_G-sHGzGq8nwflzSq8o7Gv1hWuQ,1382
+protein_quest/emdb.py,sha256=QEeU0VJQ4lLM-o5yAU3QZlrtzDZNgnC5fCjlqPtTyAY,1370
+protein_quest/filters.py,sha256=-gasSXR4g5SzYSYbkfcDwR-tm2KCAhCMdpIVJrUPR1w,5224
+protein_quest/go.py,sha256=lZNEcw8nTc9wpV3cl4y2FG9Lsj8wsXQ6zemmAQs_DWE,5650
+protein_quest/mcp_server.py,sha256=auftrx4aBZp1P-pBcunkPiSmXLtOIZ6MTuhUuW7yrGY,7241
+protein_quest/parallel.py,sha256=ZJrLO1t2HXs4EeNctytvBTyROPBq-4-gLf35PiolHf0,3468
+protein_quest/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
+protein_quest/ss.py,sha256=MMHgqKPxjYpjyExiqslWjmyG7aeForeAeJorCYdh75g,9663
+protein_quest/taxonomy.py,sha256=4mKv8zll4mX02Ow8CTvyqMJE2KJZvcq3QlTjjjLOJJk,5072
+protein_quest/uniprot.py,sha256=8qWV4GWqHTRfed0bE_TdgsLYcnDT_vzKu-6JxIgapJQ,18680
+protein_quest/utils.py,sha256=z4PPPcog6nvPhA93DWVf7stv5uJ4h_2BP5owdhoO5mo,5626
+protein_quest/alphafold/__init__.py,sha256=Ktasi5BRp71wO7-PpOGDpIRRtBEefs8knIdlKQeLQpk,51
+protein_quest/alphafold/confidence.py,sha256=pYIuwYdkuPuHLagcX1dSvSyZ_84xboRLfHUxkEoc4MY,6766
+protein_quest/alphafold/entry_summary.py,sha256=GtE3rT7wH3vIOOeiXY2s80Fo6EzdoqlcvakW8K591Yk,1257
+protein_quest/alphafold/fetch.py,sha256=iFHORaO-2NvPwmpm33tfOFUcSJx8mBGwMXxwc4bRuk8,11336
+protein_quest/pdbe/__init__.py,sha256=eNNHtN60NAGea7gvRkIzkoTXsYPK99s-ldIcKWYO6So,61
+protein_quest/pdbe/fetch.py,sha256=tlCrWoaOrwxnQFrf-PnimUUa6lmtHwwysS51efYsBcA,2379
+protein_quest/pdbe/io.py,sha256=iGLvmsD-eEYnrgZDYfkGWIDCzwDRRD5dwqB480talCs,10037
+protein_quest-0.3.2.dist-info/METADATA,sha256=wcURSjBlmkCt-ddhZX7xRYrL-7tT1VuBpJ36_mP0Iuk,7760
+protein_quest-0.3.2.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
+protein_quest-0.3.2.dist-info/entry_points.txt,sha256=f1RtOxv9TFBO3w01EMEuFXBTMsqKsQcKlkxmj9zE-0g,57
+protein_quest-0.3.2.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
+protein_quest-0.3.2.dist-info/RECORD,,

protein_quest-0.3.0.dist-info/RECORD

@@ -1,24 +0,0 @@
-protein_quest/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-protein_quest/__version__.py,sha256=VrXpHDu3erkzwl_WXrqINBm9xWkcyUy53IQOj042dOs,22
-protein_quest/cli.py,sha256=oyDin6Z92Q17mUmTCasKgju3YUJbPu298gniNakQUwY,31121
-protein_quest/emdb.py,sha256=QEeU0VJQ4lLM-o5yAU3QZlrtzDZNgnC5fCjlqPtTyAY,1370
-protein_quest/filters.py,sha256=GNtM1N1S1mNUqAvX7OvyhOvnUWo4qx2hMneORbc-Qz8,3797
-protein_quest/go.py,sha256=ycV3-grxuIKFt28bFgH6iRKmt5AEGi7txoTbaAnBxQE,5684
-protein_quest/mcp_server.py,sha256=xIaOy6sY_gW5R_oMImI2yBmbBGtZZICOxXLzOkFmm-w,7197
-protein_quest/parallel.py,sha256=kCH6KCJYJZVoq0_Qz8ZLbHnf2OJG-h4uxd9oH2rLNKc,2201
-protein_quest/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
-protein_quest/taxonomy.py,sha256=kAKKZT_mOtmX8ZWNIE9i7emE23VEewkj12X7d_t3p2Y,4659
-protein_quest/uniprot.py,sha256=8qWV4GWqHTRfed0bE_TdgsLYcnDT_vzKu-6JxIgapJQ,18680
-protein_quest/utils.py,sha256=HUvqfsuMBIFOVFlb_QC2to_UQkiZ0_fwHLlckifuXss,3700
-protein_quest/alphafold/__init__.py,sha256=Ktasi5BRp71wO7-PpOGDpIRRtBEefs8knIdlKQeLQpk,51
-protein_quest/alphafold/confidence.py,sha256=-lbwijzVMhRd98bxwFDbSi7idiUKJ5BpOsGFrvuTEnQ,5596
-protein_quest/alphafold/entry_summary.py,sha256=P-S8qrXkU-wwIccA1nGol1lfDkUW0Sg0th_3EU-WjN8,1187
-protein_quest/alphafold/fetch.py,sha256=eq__PfqisuUIQBUM8KVghpiEOBGF-zXWNC6Ll_Hlz2E,11828
-protein_quest/pdbe/__init__.py,sha256=eNNHtN60NAGea7gvRkIzkoTXsYPK99s-ldIcKWYO6So,61
-protein_quest/pdbe/fetch.py,sha256=iTyS4ucV2KZl4jTgrUFOZhsXs3cRUIuvmTbXNm_pY8U,1850
-protein_quest/pdbe/io.py,sha256=0ldsrIHKaaurrM2FfWXbqm1iRj3q6xw8-lptfYU1yEw,6231
-protein_quest-0.3.0.dist-info/METADATA,sha256=yiHZn4gDdwilbCoxrF0pCjVk04v_O5pwpwrtr6oPLrE,7369
-protein_quest-0.3.0.dist-info/WHEEL,sha256=qtCwoSJWgHk21S1Kb4ihdzI2rlJ1ZKaIurTj_ngOhyQ,87
-protein_quest-0.3.0.dist-info/entry_points.txt,sha256=f1RtOxv9TFBO3w01EMEuFXBTMsqKsQcKlkxmj9zE-0g,57
-protein_quest-0.3.0.dist-info/licenses/LICENSE,sha256=xx0jnfkXJvxRnG63LTGOxlggYnIysveWIZ6H3PNdCrQ,11357
-protein_quest-0.3.0.dist-info/RECORD,,