pxmeter 0.1.0__tar.gz → 0.1.2__tar.gz

pxmeter-0.1.2/PKG-INFO

@@ -0,0 +1,32 @@
+Metadata-Version: 2.4
+Name: pxmeter
+Version: 0.1.2
+Summary: PXMeter is a comprehensive toolkit for evaluating the quality of         structures generated by biomolecular structure prediction models.
+Author: Bytedance Inc.
+Author-email: ai4s-bio@bytedance.com
+License: Apache 2.0 License
+Platform: manylinux1
+Requires-Python: >=3.11
+License-File: LICENSE
+Requires-Dist: biotite>=1.2.0
+Requires-Dist: dockq==2.1.3
+Requires-Dist: gemmi==0.7.0
+Requires-Dist: joblib
+Requires-Dist: ml_collections
+Requires-Dist: numpy
+Requires-Dist: pandas
+Requires-Dist: pdbeccdutils==0.8.5
+Requires-Dist: posebusters==0.3.1
+Requires-Dist: rdkit>=2024.09.6
+Requires-Dist: scipy
+Requires-Dist: tabulate
+Requires-Dist: tqdm
+Requires-Dist: click
+Dynamic: author
+Dynamic: author-email
+Dynamic: license
+Dynamic: license-file
+Dynamic: platform
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary

{pxmeter-0.1.0 → pxmeter-0.1.2}/README.md

@@ -4,8 +4,8 @@
 [![Python 3.11+](https://img.shields.io/badge/python-3.11%2B-blue.svg)](https://www.python.org/downloads/)
 
 <div align="left" style="margin: 20px 0;">
-<span style="margin: 0 10px;">📄 <a href="URL">From Dataset Diagnostics to Evaluation: Revisiting
-Structure Prediction Benchmarks with PXMeter</a></span>
+<span style="margin: 0 10px;">📄 <a href="https://www.biorxiv.org/content/10.1101/2025.07.17.664878v1">From Dataset Curation to Unified Evaluation: Revisiting
+ Structure Prediction Benchmarks with PXMeter</a></span>
 </div>
 
 PXMeter is a comprehensive toolkit for evaluating the quality of structures generated by biomolecular structure prediction models, with support for proteins, nucleic acids, and small molecules.
@@ -86,7 +86,12 @@ Refer to [benchmark/README.md](./benchmark/README.md) for evaluation protocols o
 - PoseBusters V2
 
 The benchmark data is released under the CC0 license.
-We include code in the `benchmark` directory that evaluates various models using PXMeter and aggregates their metrics. This serves as an example of best practices for using the tool. For more details, please refer to the [Technical Report URL].
+We include code in the `benchmark` directory that evaluates various models using PXMeter and aggregates their metrics.
+This serves as an example of best practices for using the tool. For more details, please refer to our paper:
+
+📄 <a href="https://www.biorxiv.org/content/10.1101/2025.07.17.664878v1">From Dataset Curation to Unified Evaluation: Revisiting
+ Structure Prediction Benchmarks with PXMeter</a>
+
 
 
 ## 💪 Contributing to PXMeter
@@ -102,21 +107,27 @@ pre-commit install
 ```
 
 
-## 🚧 Limitations
-- PXMeter supports chain/atom permutations but not residue-level permutations. As a result, the accuracy of evaluation for branched chains, such as glycans, cannot be fully guaranteed.
-- It is recommended to use CIF files from the RCSB PDB as references, as they ensure content accuracy. All development and testing were conducted exclusively on CIF files from this source.
-
-
 ## ✍️ Citing PXMeter
+If you use PXMeter in your research, please cite the following:
 ```bibtex
-@software{pxmeter,
-  title = {PXMeter: Biomolecular Structure Quality Assessment Toolkit},
-  author = {Bytedance AI for Science Team},
-  year = {2024},
-  url = {https://github.com/your-org/pxmeter},
+@article {Ma2025.07.17.664878,
+	author = {Ma, Wenzhi and Liu, Zhenyu and Yang, Jincai and Lu, Chan and Zhang, Hanyu and Xiao, Wenzhi},
+	title = {From Dataset Curation to Unified Evaluation: Revisiting Structure Prediction Benchmarks with PXMeter},
+	year = {2025},
+	doi = {10.1101/2025.07.17.664878},
+	publisher = {Cold Spring Harbor Laboratory},
+	URL = {https://www.biorxiv.org/content/early/2025/07/22/2025.07.17.664878},
+	eprint = {https://www.biorxiv.org/content/early/2025/07/22/2025.07.17.664878.full.pdf},
+	journal = {bioRxiv}
 }
 ```
 
+
+## 🚧 Limitations
+- PXMeter supports chain/atom permutations but not residue-level permutations. As a result, the accuracy of evaluation for branched chains, such as glycans, cannot be fully guaranteed.
+- It is recommended to use CIF files from the RCSB PDB as references, as they ensure content accuracy. All development and testing were conducted exclusively on CIF files from this source.
+
+
 ## 🛡️ Security
 If you discover a potential security issue in this project, or think you may
 have discovered a security issue, we ask that you notify Bytedance Security via our [security center](https://security.bytedance.com/src) or [vulnerability reporting email](sec@bytedance.com).

{pxmeter-0.1.0 → pxmeter-0.1.2}/pxmeter/cli.py

@@ -18,9 +18,10 @@ from pathlib import Path
 import click
 
 from pxmeter.configs.data_config import (
+    CCD_BLOCKS_JSON,
     COMPONENTS_FILE,
     download_ccd_cif,
-    make_one_letter_code_json_from_ccd,
+    make_precomputed_json_from_ccd,
     ONE_LETTER_CODE_JSON,
 )
 from pxmeter.eval import evaluate, MetricResult
@@ -176,4 +177,6 @@ def update():
     Update the CCD database.
     """
     download_ccd_cif(output_path=COMPONENTS_FILE.parent)
-    make_one_letter_code_json_from_ccd(COMPONENTS_FILE, ONE_LETTER_CODE_JSON)
+    make_precomputed_json_from_ccd(
+        COMPONENTS_FILE, CCD_BLOCKS_JSON, ONE_LETTER_CODE_JSON
+    )

{pxmeter-0.1.0 → pxmeter-0.1.2}/pxmeter/configs/data_config.py

@@ -55,7 +55,9 @@ def download_ccd_cif(output_path: Path):
     logging.info("Download CCD CIF file successfully: %s", output_cif)
 
 
-def make_one_letter_code_json_from_ccd(components_file: Path, output_json: Path):
+def make_precomputed_json_from_ccd(
+    components_file: Path, ccd_blocks_json: Path, one_letter_code_json: Path
+):
     """
     Make a one-letter code JSON file from the CCD CIF file.
 
@@ -65,17 +67,23 @@ def make_one_letter_code_json_from_ccd(components_file: Path, output_json: Path)
     ccd_cif = gemmi.cif.read(str(components_file))
 
     ccd_code_to_one_letter_code = {}
+    ccd_code_to_block = {}
     for block in ccd_cif:
         ccd_code = block.find_value("_chem_comp.id")
         one_letter_code = block.find_value("_chem_comp.one_letter_code")
+        ccd_code_to_block[ccd_code] = block.as_string()
         if one_letter_code is None or one_letter_code == "?":
             continue
         ccd_code_to_one_letter_code[ccd_code] = one_letter_code
 
-    with open(output_json, "w") as f:
-        json.dump(ccd_code_to_one_letter_code, f, indent=4)
+    with open(ccd_blocks_json, "w") as f:
+        json.dump(ccd_code_to_block, f, indent=4)
+
+    logging.info("Make CCD_BLOCKS_JSON successfully: %s", ccd_blocks_json)
 
-    logging.info("Make ONE_LETTER_CODE_JSON successfully: %s", output_json)
+    with open(one_letter_code_json, "w") as f:
+        json.dump(ccd_code_to_one_letter_code, f, indent=4)
+    logging.info("Make ONE_LETTER_CODE_JSON successfully: %s", one_letter_code_json)
 
 
 # default is <repo_dir>/ccd_cache/components.cif Your path for components file
@@ -84,20 +92,30 @@ def make_one_letter_code_json_from_ccd(components_file: Path, output_json: Path)
 repo_dir = Path(__file__).absolute().parent.parent.parent
 ccd_file_in_repo = repo_dir / "ccd_cache" / "components.cif"
 COMPONENTS_FILE = Path(os.environ.get("PXM_CCD_FILE", ccd_file_in_repo))
-ONE_LETTER_CODE_JSON = COMPONENTS_FILE.with_suffix(".json")
+
+CCD_BLOCKS_JSON = COMPONENTS_FILE.with_suffix(".json")
+ONE_LETTER_CODE_JSON = COMPONENTS_FILE.parent / "one_letter_code.json"
 
 if not COMPONENTS_FILE.exists():
     logging.debug(
         "CCD CIF file not found. Downloading CCD CIF file to %s", COMPONENTS_FILE.parent
     )
     download_ccd_cif(output_path=COMPONENTS_FILE.parent)
-    make_one_letter_code_json_from_ccd(COMPONENTS_FILE, ONE_LETTER_CODE_JSON)
+    make_precomputed_json_from_ccd(
+        COMPONENTS_FILE, CCD_BLOCKS_JSON, ONE_LETTER_CODE_JSON
+    )
 else:
     logging.debug("Load CCD CIF file from: %s", COMPONENTS_FILE)
 
-if not ONE_LETTER_CODE_JSON.exists():
-    make_one_letter_code_json_from_ccd(COMPONENTS_FILE, ONE_LETTER_CODE_JSON)
+if not ONE_LETTER_CODE_JSON.exists() or not CCD_BLOCKS_JSON.exists():
+    make_precomputed_json_from_ccd(
+        COMPONENTS_FILE, CCD_BLOCKS_JSON, ONE_LETTER_CODE_JSON
+    )
 
 logging.debug("Load CCD one-letter code from: %s", ONE_LETTER_CODE_JSON)
 with open(ONE_LETTER_CODE_JSON, "r") as f:
     CCD_ONE_LETTER_CODE = json.load(f)
+
+
+with open(CCD_BLOCKS_JSON, "r") as f:
+    CCD_BLOCKS = json.load(f)

{pxmeter-0.1.0 → pxmeter-0.1.2}/pxmeter/data/ccd.py

@@ -24,7 +24,7 @@ from pdbeccdutils.core import ccd_reader
 from rdkit import Chem
 from rdkit.Geometry import Point3D
 
-from pxmeter.configs.data_config import COMPONENTS_FILE
+from pxmeter.configs.data_config import CCD_BLOCKS
 from pxmeter.data.utils import (
     get_inter_residue_bonds,
     get_mol_graph_matches,
@@ -34,17 +34,6 @@ from pxmeter.data.utils import (
 logging.getLogger("rdkit").setLevel(logging.ERROR)
 
 
-@functools.lru_cache
-def _gemmi_load_ccd_cif() -> gemmi.cif.Document:
-    """
-    Load the CCD components file using the gemmi library.
-
-    Returns:
-        gemmi.cif.Document: gemmi ccd components file
-    """
-    return gemmi.cif.read(str(COMPONENTS_FILE))
-
-
 @functools.lru_cache
 def get_ccd_mol_from_cif(ccd_code: str) -> Chem.Mol:
     """
@@ -57,10 +46,8 @@ def get_ccd_mol_from_cif(ccd_code: str) -> Chem.Mol:
         mol (Chem.Mol): The RDKit molecule object corresponding to the given CCD code.
                               Returns None if the CCD code is not found.
     """
-    ccd_cif = _gemmi_load_ccd_cif()
-
     try:
-        ccd_block = ccd_cif[ccd_code]
+        ccd_block = gemmi.cif.read_string(CCD_BLOCKS[ccd_code])[0]
     except KeyError:
         return
     ccd_reader_result = ccd_reader._parse_pdb_mmcif(ccd_block, sanitize=True)
@@ -232,7 +219,7 @@ def get_ccd_perm_info(ccd_code: str) -> dict[str, Any]:
         for node in removed_nodes:
             mol_graph.remove_node(node)
 
-        matches = get_mol_graph_matches(mol_graph, mol_graph)[:1000]
+        matches = get_mol_graph_matches(mol_graph, mol_graph, max_matches=1000)
 
         # re-index after removing H
         reverted_old_atom_map = {v: k for k, v in mol.atom_map.items()}

{pxmeter-0.1.0 → pxmeter-0.1.2}/pxmeter/data/utils.py

@@ -234,13 +234,16 @@ def rdkit_mol_to_nx_graph(mol: Chem.Mol) -> nx.Graph:
     return G
 
 
-def get_mol_graph_matches(mol_graph1: nx.Graph, mol_graph2: nx.Graph) -> list[dict]:
+def get_mol_graph_matches(
+    mol_graph1: nx.Graph, mol_graph2: nx.Graph, max_matches: int = 1000
+) -> list[dict]:
     """
     Find all isomorphisms between subgraph of mol_graph1 and mol_graph2.
 
     Args:
         mol_graph1 (nx.Graph): Source molecular graph (typically larger subgraph)
         mol_graph2 (nx.Graph): Target molecular graph (typically smaller supergraph)
+        max_matches (int): Maximum number of matches to return. Default is 1000.
 
     Returns:
         list[dict]: List of node mapping dictionaries where each dictionary represents
@@ -254,4 +257,12 @@ def get_mol_graph_matches(mol_graph1: nx.Graph, mol_graph2: nx.Graph) -> list[di
         mol_graph2,
         node_match=lambda x, y: x["atomic_num"] == y["atomic_num"],
     )
-    return list(isomatcher.subgraph_isomorphisms_iter())
+
+    matches = []
+    num = 0
+    for i in isomatcher.subgraph_isomorphisms_iter():
+        matches.append(i)
+        num += 1
+        if num >= max_matches:
+            break
+    return matches

{pxmeter-0.1.0 → pxmeter-0.1.2}/pxmeter/mapping.py

@@ -584,7 +584,7 @@ class MappingCIF:
             ref_graph = rdkit_mol_to_nx_graph(ref_mol)
             model_graph = rdkit_mol_to_nx_graph(model_mol)
             try:
-                matches = get_mol_graph_matches(model_graph, ref_graph)
+                matches = get_mol_graph_matches(model_graph, ref_graph, max_matches=1)
                 assert len(matches) > 0
                 match = matches[0]
 

pxmeter-0.1.2/pxmeter.egg-info/PKG-INFO

@@ -0,0 +1,32 @@
+Metadata-Version: 2.4
+Name: pxmeter
+Version: 0.1.2
+Summary: PXMeter is a comprehensive toolkit for evaluating the quality of         structures generated by biomolecular structure prediction models.
+Author: Bytedance Inc.
+Author-email: ai4s-bio@bytedance.com
+License: Apache 2.0 License
+Platform: manylinux1
+Requires-Python: >=3.11
+License-File: LICENSE
+Requires-Dist: biotite>=1.2.0
+Requires-Dist: dockq==2.1.3
+Requires-Dist: gemmi==0.7.0
+Requires-Dist: joblib
+Requires-Dist: ml_collections
+Requires-Dist: numpy
+Requires-Dist: pandas
+Requires-Dist: pdbeccdutils==0.8.5
+Requires-Dist: posebusters==0.3.1
+Requires-Dist: rdkit>=2024.09.6
+Requires-Dist: scipy
+Requires-Dist: tabulate
+Requires-Dist: tqdm
+Requires-Dist: click
+Dynamic: author
+Dynamic: author-email
+Dynamic: license
+Dynamic: license-file
+Dynamic: platform
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary

{pxmeter-0.1.0 → pxmeter-0.1.2}/setup.py

@@ -20,7 +20,7 @@ with open("requirements.txt") as f:
 setup(
     name="pxmeter",
     python_requires=">=3.11",
-    version="0.1.0",
+    version="0.1.2",
     description="PXMeter is a comprehensive toolkit for evaluating the quality of \
         structures generated by biomolecular structure prediction models.",
     author="Bytedance Inc.",

pxmeter-0.1.0/PKG-INFO

@@ -1,10 +0,0 @@
-Metadata-Version: 2.1
-Name: pxmeter
-Version: 0.1.0
-Summary: PXMeter is a comprehensive toolkit for evaluating the quality of         structures generated by biomolecular structure prediction models.
-Author: Bytedance Inc.
-Author-email: ai4s-bio@bytedance.com
-License: Apache 2.0 License
-Platform: manylinux1
-Requires-Python: >=3.11
-License-File: LICENSE

pxmeter-0.1.0/pxmeter.egg-info/PKG-INFO

@@ -1,10 +0,0 @@
-Metadata-Version: 2.1
-Name: pxmeter
-Version: 0.1.0
-Summary: PXMeter is a comprehensive toolkit for evaluating the quality of         structures generated by biomolecular structure prediction models.
-Author: Bytedance Inc.
-Author-email: ai4s-bio@bytedance.com
-License: Apache 2.0 License
-Platform: manylinux1
-Requires-Python: >=3.11
-License-File: LICENSE