Add namespace and metadata information

src/pathme/kegg/convert_to_bel.py

@@ -18,6 +18,7 @@
     get_all_reactions, get_all_relationships, get_complex_components, get_entity_nodes, get_reaction_pathway_edges,
     import_xml_etree,
 )
+from pathme.utils import add_bel_metadata
 
 __all__ = [
     'kegg_to_bel',
@@ -49,6 +50,8 @@ def kegg_to_bel(path, hgnc_manager, chebi_manager, flatten=False):
         contact='daniel.domingo.fernandez@scai.fraunhofer.de'
     )
 
+    add_bel_metadata(graph)
+
     graph.graph['pathway_id'] = root.attrib['name']
 
     # Parse file and get entities and interactions

src/pathme/reactome/convert_to_bel.py

@@ -24,7 +24,7 @@
 
 from pathme.constants import UNKNOWN, REACTOME_CITATION
 from pathme.reactome.utils import get_valid_node_parameters, process_multiple_proteins
-from pathme.utils import parse_id_uri
+from pathme.utils import add_bel_metadata, parse_id_uri
 
 log = logging.getLogger(__name__)
 
@@ -56,6 +56,8 @@ def convert_to_bel(nodes: Dict[str, Dict], interactions: List[Tuple[str, str, Di
         contact='daniel.domingo.fernandez@scai.fraunhofer.de',
     )
 
+    add_bel_metadata(graph)
+
     graph.graph['pathway_id'] = identifier
 
     nodes = nodes_to_bel(nodes, graph, hgnc_manager, chebi_manager)

src/pathme/utils.py

@@ -5,7 +5,6 @@
 import collections
 import itertools as itt
 import logging
-import os
 import pickle
 from typing import Dict, Iterable, List, Optional, Set, Tuple
 from urllib.parse import urlparse
@@ -16,9 +15,10 @@
 import pybel
 import rdflib
 from pybel import BELGraph, from_pickle
+from pybel.constants import GRAPH_NAMESPACE_URL
 from pybel.struct.summary import count_functions, count_relations
 
-from pathme.constants import BEL_STATS_COLUMN_NAMES, UNKNOWN
+from pathme.constants import *
 from pathme.export_utils import get_paths_in_folder
 
 log = logging.getLogger(__name__)
@@ -515,3 +515,31 @@ def summarize_helper(graphs: Iterable[BELGraph]):
     summary_str = graph.summary_str()
 
     click.echo(summary_str)
+
+
+def add_bel_metadata(graph: BELGraph) -> None:
+    """Add namespaces and annotations to a given bel graph."""
+    graph.graph[GRAPH_NAMESPACE_URL] = {
+        CHEBI.upper(): "https://raw.githubusercontent.com/pharmacome/terminology/b46b65c3da259b6e86026514dfececab7c22a11b/external/chebi-names.belns",
+        HGNC: "https://raw.githubusercontent.com/pharmacome/terminology/b46b65c3da259b6e86026514dfececab7c22a11b/external/hgnc-names.belns",
+        "GO": "https://raw.githubusercontent.com/pharmacome/terminology/b46b65c3da259b6e86026514dfececab7c22a11b/external/go-names.belns",
+    }
+
+    graph.namespace_pattern["NCBIGENE"] = "^\d+$"
+    graph.namespace_pattern[UNIPROT.upper()] = \
+        "^([A-N,R-Z][0-9]([A-Z][A-Z, 0-9][A-Z, 0-9][0-9]){1,2})|([O,P,Q][0-9][A-Z, 0-9][A-Z, 0-9][A-Z, 0-9][0-9])(\.\d+)?$"
+    graph.namespace_pattern[PUBCHEM.upper()] = ".*"
+    graph.namespace_pattern[EXPASY] = ".*"
+    graph.namespace_pattern[ENTREZ] = ".*"
+    graph.namespace_pattern[ENSEMBL] = ".*"
+    graph.namespace_pattern[WIKIPEDIA] = ".*"
+    graph.namespace_pattern[MIRBASE.upper()] = ".*"
+    graph.namespace_pattern[INTERPRO.upper()] = ".*"
+    graph.namespace_pattern[PFAM.upper()] = ".*"
+    graph.namespace_pattern[BRENDA.upper()] = ".*"
+
+    graph.namespace_pattern[KEGG.upper()] = ".*"
+    graph.namespace_pattern[REACTOME.upper()] = ".*"
+    graph.namespace_pattern[WIKIPATHWAYS.upper()] = ".*"
+
+    graph.annotation_list['database'] = {KEGG, REACTOME, WIKIPATHWAYS}

src/pathme/wikipathways/convert_to_bel.py

@@ -12,7 +12,7 @@
 
 from .utils import check_multiple, evaluate_wikipathways_metadata, get_valid_gene_identifier
 from ..constants import HGNC
-from ..utils import parse_id_uri
+from ..utils import add_bel_metadata, parse_id_uri
 
 log = logging.getLogger(__name__)
 
@@ -48,6 +48,8 @@ def convert_to_bel(
         contact='daniel.domingo.fernandez@scai.fraunhofer.de',
     )
 
+    add_bel_metadata(graph)
+
     pathway_id = graph.graph['pathway_id'] = pathway_info['pathway_id']
 
     nodes = {