diff --git a/translate2modelcif.py b/translate2modelcif.py
index 77f93e71f17cf605017b762c1e86d9659f739a6b..3d87f9034a09bdab8fd0581cbc3652fe91183526 100644
--- a/translate2modelcif.py
+++ b/translate2modelcif.py
@@ -755,43 +755,64 @@ def _get_sequence_dbs(seq_dbs):
return [db_dict[seq_db] for seq_db in seq_dbs]
+def _get_modelcif_protocol_software(js_step):
+ """Assemble software entries for a ModelCIF protocol step."""
+ if js_step["software"]:
+ if len(js_step["software"]) == 1:
+ sftwre = _assemble_modelcif_software(js_step["software"][0])
+ else:
+ sftwre = []
+ for sft in js_step["software"]:
+ sftwre.append(_assemble_modelcif_software(sft))
+ sftwre = modelcif.SoftwareGroup(elements=sftwre)
+ if js_step["software_parameters"]:
+ params = []
+ for key, val in js_step["software_parameters"].items():
+ params.append(modelcif.SoftwareParameter(key, val))
+ if isinstance(sftwre, modelcif.SoftwareGroup):
+ sftwre.parameters = params
+ else:
+ sftwre = modelcif.SoftwareGroup(
+ elements=(sftwre,), parameters=params
+ )
+ return sftwre
+
+ return None
+
+
+def _get_modelcif_protocol_input(js_step, target_entities, ref_dbs, model):
+ """Assemble input data for a ModelCIF protocol step."""
+ if js_step["input"] == "target_sequences":
+ input_data = modelcif.data.DataGroup(target_entities)
+ input_data.extend(ref_dbs)
+ elif js_step["input"] == "model":
+ input_data = model
+ else:
+ raise RuntimeError(f"Unknown protocol input: '{js_step['input']}'")
+
+ return input_data
+
+
+def _get_modelcif_protocol_output(js_step, model):
+ """Assemble output data for a ModelCIF protocol step."""
+ if js_step["output"] == "model":
+ output_data = model
+ else:
+ raise RuntimeError(f"Unknown protocol output: '{js_step['output']}'")
+
+ return output_data
+
+
def _get_modelcif_protocol(protocol_steps, target_entities, model, ref_dbs):
"""Create the protocol for the ModelCIF file."""
protocol = modelcif.protocol.Protocol()
for js_step in protocol_steps:
- sftwre = None
- if js_step["software"]:
- if len(js_step["software"]) == 1:
- sftwre = _assemble_modelcif_software(js_step["software"][0])
- else:
- sftwre = []
- for sft in js_step["software"]:
- sftwre.append(_assemble_modelcif_software(sft))
- sftwre = modelcif.SoftwareGroup(elements=sftwre)
- if js_step["software_parameters"]:
- params = []
- for k, v in js_step["software_parameters"].items():
- params.append(modelcif.SoftwareParameter(k, v))
- if isinstance(sftwre, modelcif.SoftwareGroup):
- sftwre.parameters = params
- else:
- sftwre = modelcif.SoftwareGroup(
- elements=(sftwre,), parameters=params
- )
+ sftwre = _get_modelcif_protocol_software(js_step)
+ input_data = _get_modelcif_protocol_input(
+ js_step, target_entities, ref_dbs, model
+ )
+ output_data = _get_modelcif_protocol_output(js_step, model)
- if js_step["input"] == "target_sequences":
- input_data = modelcif.data.DataGroup(target_entities)
- input_data.extend(ref_dbs)
- elif js_step["input"] == "model":
- input_data = model
- else:
- raise RuntimeError(f"Unknown protocol input: '{js_step['input']}'")
- if js_step["output"] == "model":
- output_data = model
- else:
- raise RuntimeError(
- f"Unknown protocol output: '{js_step['output']}'"
- )
protocol.steps.append(
modelcif.protocol.Step(
input_data=input_data,
@@ -826,33 +847,31 @@ def _package_associated_files(mdl_name):
os.remove(file)
-def _store_as_modelcif(
- interaction_name, data_json, ost_ent, out_dir, file_prfx, compress
-):
+def _store_as_modelcif(data_json, ost_ent, out_dir, file_prfx, compress):
"""Mix all the data into a ModelCIF file."""
print(" generating ModelCIF objects...", end="")
pstart = timer()
# create system to gather all the data
system = modelcif.System(
title=data_json["title"],
- id=interaction_name.upper(),
+ id=data_json["data_block_id"].upper(),
model_details=data_json["model_details"],
)
+
# create target entities, references, source, asymmetric units & assembly
# for source we assume all chains come from the same taxon
- source = ihm.source.Natural(
- ncbi_taxonomy_id=data_json["target_entities"][0]["up_ncbi_taxid"],
- scientific_name=data_json["target_entities"][0]["up_organism"],
- )
-
# create an asymmetric unit and an entity per target sequence
asym_units = {}
_get_modelcif_entities(
- data_json["target_entities"], source, asym_units, system
+ data_json["target_entities"],
+ ihm.source.Natural(
+ ncbi_taxonomy_id=data_json["target_entities"][0]["up_ncbi_taxid"],
+ scientific_name=data_json["target_entities"][0]["up_organism"],
+ ),
+ asym_units,
+ system,
)
- assembly = modelcif.Assembly(asym_units.values())
-
# audit_authors
system.authors.extend(data_json["audit_authors"])
@@ -865,7 +884,7 @@ def _store_as_modelcif(
f"(#{data_json['rank_num']} ranked model)"
)
model = _OST2ModelCIF(
- assembly=assembly,
+ assembly=modelcif.Assembly(asym_units.values()),
asym=asym_units,
ost_entity=ost_ent,
name=mdl_list_name,
@@ -877,10 +896,9 @@ def _store_as_modelcif(
system.repositories.append(model.add_scores(data_json, system.id, mdl_name))
print(f" ({timer()-pstart:.2f}s)")
- model_group = modelcif.model.ModelGroup(
- [model], name=data_json["model_group_name"]
+ system.model_groups.append(
+ modelcif.model.ModelGroup([model], name=data_json["model_group_name"])
)
- system.model_groups.append(model_group)
ref_dbs = _get_sequence_dbs(data_json["config_data"]["seq_dbs"])
protocol = _get_modelcif_protocol(
@@ -919,7 +937,7 @@ def _create_interaction_json(config_data):
return data
-def _create_model_json(data, pdb_file, up_acs):
+def _create_model_json(data, pdb_file, up_acs, block_id):
"""Create a dictionary (mimicking JSON) that contains all the data."""
data["target_entities"], ost_ent = _get_entities(pdb_file, up_acs)
@@ -927,6 +945,7 @@ def _create_model_json(data, pdb_file, up_acs):
for i in data["target_entities"]:
gns.append(i["up_gn"])
data["title"] = _get_title(gns)
+ data["data_block_id"] = block_id
data["model_details"] = _get_model_details(gns)
data["model_group_name"] = _get_model_group_name()
@@ -964,8 +983,6 @@ def _main():
print(" preparing data...", end="")
pstart = timer()
- # NOTE: could also be prepared globally if all carefully overwritten
- # but not worth the trouble...
mdlcf_json = _create_interaction_json(config_data)
# uid = ..._rank_X_model_Y.pdb
@@ -975,14 +992,14 @@ def _main():
mdlcf_json["rank_num"] = int(mdl_name_parts[-3])
mdlcf_json["mdl_num"] = int(mdl_name_parts[-1])
- ost_ent = _create_model_json(mdlcf_json, fle, up_acs)
+ ost_ent = _create_model_json(mdlcf_json, fle, up_acs, uid)
# read quality scores from JSON file
_get_scores(mdlcf_json, file_prfx)
print(f" ({timer()-pstart:.2f}s)")
_store_as_modelcif(
- uid, mdlcf_json, ost_ent, opts.out_dir, file_prfx, opts.compress
+ mdlcf_json, ost_ent, opts.out_dir, file_prfx, opts.compress
)
print(f" ... done with {fle} ({timer()-pdb_start:.2f}s).")