nmdc-schema 11.12.0rc2__py3-none-any.whl → 11.12.1__py3-none-any.whl

nmdc_schema/nmdc.schema.json

@@ -2647,9 +2647,9 @@
                     ]
                 },
                 "misc_param": {
-                    "description": "Any other measurement performed or parameter collected, that is not listed here",
+                    "description": "Structured miscellaneous property assertions for this Biosample. Use when a value cannot cleanly fit an existing, policy-governed slot.",
                     "items": {
-                        "$ref": "#/$defs/TextValue"
+                        "$ref": "#/$defs/PropertyAssertion"
                     },
                     "type": [
                         "array",
@@ -8515,7 +8515,7 @@
         },
         "Instrument": {
             "additionalProperties": false,
-            "description": "A material entity that is designed to perform a function in a scientific investigation, but is not a reagent.",
+            "description": "A material entity that is designed to perform a function in a scientific investigation, but is not a reagent. This class models the make and model of the instrument, not the specific instance of the instrument.",
             "properties": {
                 "alternative_identifiers": {
                     "description": "A list of alternative identifiers for the entity.",
@@ -13128,10 +13128,113 @@
             "title": "ProfilePositionEnum",
             "type": "string"
         },
+        "PropertyAssertion": {
+            "additionalProperties": false,
+            "description": "A structured record of data that doesn't fit nicely within the constraints of other NMDC AttributeValues. Uses primitive ranges only.",
+            "properties": {
+                "has_attribute_id": {
+                    "description": "CURIE or IRI for the property (MIxS slot, NMDC slot, ENVO/OBI term, etc.).",
+                    "type": [
+                        "string",
+                        "null"
+                    ]
+                },
+                "has_attribute_label": {
+                    "description": "Human-readable label for the property (e.g., MIxS label, ENVO term label).",
+                    "type": [
+                        "string",
+                        "null"
+                    ]
+                },
+                "has_boolean_value": {
+                    "description": "Links a quantity value to a boolean",
+                    "type": [
+                        "boolean",
+                        "null"
+                    ]
+                },
+                "has_datetime_value": {
+                    "description": "Date-time value for the property in ISO-8601 format.",
+                    "type": [
+                        "string",
+                        "null"
+                    ]
+                },
+                "has_maximum_numeric_value": {
+                    "description": "The maximum value part, expressed as number, of the quantity value when the value covers a range.",
+                    "type": [
+                        "number",
+                        "null"
+                    ]
+                },
+                "has_minimum_numeric_value": {
+                    "description": "The minimum value part, expressed as number, of the quantity value when the value covers a range.",
+                    "type": [
+                        "number",
+                        "null"
+                    ]
+                },
+                "has_numeric_value": {
+                    "description": "Links a quantity value to a number",
+                    "type": [
+                        "number",
+                        "null"
+                    ]
+                },
+                "has_quantity_kind_id": {
+                    "description": "Optional CURIE or IRI for the physical quantity kind (e.g., qudt:QuantityKind).",
+                    "type": [
+                        "string",
+                        "null"
+                    ]
+                },
+                "has_raw_value": {
+                    "description": "Original contributor string representation (unparsed)",
+                    "type": "string"
+                },
+                "has_unit": {
+                    "description": "UCUM unit code (required only when numeric value is present)",
+                    "type": [
+                        "string",
+                        "null"
+                    ]
+                },
+                "has_value_term_id": {
+                    "description": "CURIE or IRI for categorical values (ENVO, PATO, METPO, etc.).",
+                    "type": [
+                        "string",
+                        "null"
+                    ]
+                },
+                "type": {
+                    "description": "the class_uri of the class that has been instantiated",
+                    "enum": [
+                        "nmdc:PropertyAssertion"
+                    ],
+                    "type": "string"
+                }
+            },
+            "required": [
+                "has_raw_value",
+                "type"
+            ],
+            "title": "PropertyAssertion",
+            "type": "object"
+        },
         "Protocol": {
             "additionalProperties": false,
             "description": "",
             "properties": {
+                "analysis_type": {
+                    "description": "Select all the data types associated or available for this biosample",
+                    "items": {
+                        "$ref": "#/$defs/AnalysisTypeEnum"
+                    },
+                    "type": [
+                        "array",
+                        "null"
+                    ]
+                },
                 "description": {
                     "description": "a human-readable description of a thing",
                     "type": [
@@ -13146,6 +13249,10 @@
                         "null"
                     ]
                 },
+                "protocol_for": {
+                    "$ref": "#/$defs/ProtocolForEnum",
+                    "description": "The type of planned process that the protocol describes."
+                },
                 "type": {
                     "description": "the class_uri of the class that has been instantiated",
                     "enum": [
@@ -13184,6 +13291,43 @@
             "title": "ProtocolCategoryEnum",
             "type": "string"
         },
+        "ProtocolForEnum": {
+            "description": "The permitted values for describing the type of planned process that a protocol describes.",
+            "enum": [
+                "AnnotatingWorkflow",
+                "ChemicalConversionProcess",
+                "ChromatographicSeparationProcess",
+                "CollectingBiosamplesFromSite",
+                "DataEmitterProcess",
+                "DataGeneration",
+                "DissolvingProcess",
+                "Extraction",
+                "FiltrationProcess",
+                "LibraryPreparation",
+                "MagsAnalysis",
+                "MassSpectrometry",
+                "MaterialProcessing",
+                "MetabolomicsAnalysis",
+                "MetagenomeAnnotation",
+                "MetagenomeAssembly",
+                "MetaproteomicsAnalysis",
+                "MetatranscriptomeAnnotation",
+                "MetatranscriptomeAssembly",
+                "MetatranscriptomeExpressionAnalysis",
+                "MixingProcess",
+                "NomAnalysis",
+                "NucleotideSequencing",
+                "PlannedProcess",
+                "Pooling",
+                "ReadBasedTaxonomyAnalysis",
+                "ReadQcAnalysis",
+                "StorageProcess",
+                "SubSamplingProcess",
+                "WorkflowExecution"
+            ],
+            "title": "ProtocolForEnum",
+            "type": "string"
+        },
         "QuadPosEnum": {
             "description": "",
             "enum": [
@@ -15579,5 +15723,5 @@
     },
     "title": "NMDC",
     "type": "object",
-    "version": "11.12.0rc2"
+    "version": "11.12.1"
 }

nmdc_schema/nmdc_materialized_patterns.json

@@ -6,7 +6,7 @@
     "not importing any MIxS terms where the relationship between the name (SCN) and the id isn't 1:1"
   ],
   "id": "https://w3id.org/nmdc/nmdc",
-  "version": "11.12.0rc2",
+  "version": "11.12.1",
   "license": "https://creativecommons.org/publicdomain/zero/1.0/",
   "prefixes": {
     "BFO": {
@@ -417,47 +417,6 @@
   "default_prefix": "nmdc",
   "default_range": "string",
   "types": {
-    "external_identifier": {
-      "name": "external_identifier",
-      "description": "A CURIE representing an external identifier",
-      "from_schema": "https://w3id.org/nmdc/nmdc",
-      "see_also": [
-        "https://microbiomedata.github.io/nmdc-schema/identifiers/"
-      ],
-      "typeof": "uriorcurie",
-      "uri": "xsd:anyURI",
-      "pattern": "^[a-zA-Z0-9][a-zA-Z0-9_\\.]+:[a-zA-Z0-9_][a-zA-Z0-9_\\-\\/\\.,]*$"
-    },
-    "bytes": {
-      "name": "bytes",
-      "description": "An integer value that corresponds to a size in bytes",
-      "from_schema": "https://w3id.org/nmdc/nmdc",
-      "see_also": [
-        "UO:0000233"
-      ],
-      "base": "int",
-      "uri": "xsd:long"
-    },
-    "decimal_degree": {
-      "name": "decimal_degree",
-      "description": "A decimal degree expresses latitude or longitude as decimal fractions.",
-      "from_schema": "https://w3id.org/nmdc/nmdc",
-      "see_also": [
-        "https://en.wikipedia.org/wiki/Decimal_degrees"
-      ],
-      "base": "float",
-      "uri": "xsd:decimal"
-    },
-    "language_code": {
-      "name": "language_code",
-      "description": "A language code conforming to ISO_639-1",
-      "from_schema": "https://w3id.org/nmdc/nmdc",
-      "see_also": [
-        "https://en.wikipedia.org/wiki/ISO_639-1"
-      ],
-      "base": "str",
-      "uri": "xsd:language"
-    },
     "string": {
       "name": "string",
       "description": "A character string",
@@ -708,6 +667,47 @@
       "base": "str",
       "uri": "xsd:string",
       "repr": "str"
+    },
+    "external_identifier": {
+      "name": "external_identifier",
+      "description": "A CURIE representing an external identifier",
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "see_also": [
+        "https://microbiomedata.github.io/nmdc-schema/identifiers/"
+      ],
+      "typeof": "uriorcurie",
+      "uri": "xsd:anyURI",
+      "pattern": "^[a-zA-Z0-9][a-zA-Z0-9_\\.]+:[a-zA-Z0-9_][a-zA-Z0-9_\\-\\/\\.,]*$"
+    },
+    "bytes": {
+      "name": "bytes",
+      "description": "An integer value that corresponds to a size in bytes",
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "see_also": [
+        "UO:0000233"
+      ],
+      "base": "int",
+      "uri": "xsd:long"
+    },
+    "decimal_degree": {
+      "name": "decimal_degree",
+      "description": "A decimal degree expresses latitude or longitude as decimal fractions.",
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "see_also": [
+        "https://en.wikipedia.org/wiki/Decimal_degrees"
+      ],
+      "base": "float",
+      "uri": "xsd:decimal"
+    },
+    "language_code": {
+      "name": "language_code",
+      "description": "A language code conforming to ISO_639-1",
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "see_also": [
+        "https://en.wikipedia.org/wiki/ISO_639-1"
+      ],
+      "base": "str",
+      "uri": "xsd:language"
     }
   },
   "enums": {
@@ -7502,6 +7502,106 @@
         }
       }
     },
+    "ProtocolForEnum": {
+      "name": "ProtocolForEnum",
+      "description": "The permitted values for describing the type of planned process that a protocol describes.",
+      "comments": [
+        "These are the non-abstract class descendants of PlannedProcess."
+      ],
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "permissible_values": {
+        "AnnotatingWorkflow": {
+          "text": "AnnotatingWorkflow"
+        },
+        "ChemicalConversionProcess": {
+          "text": "ChemicalConversionProcess"
+        },
+        "ChromatographicSeparationProcess": {
+          "text": "ChromatographicSeparationProcess"
+        },
+        "CollectingBiosamplesFromSite": {
+          "text": "CollectingBiosamplesFromSite"
+        },
+        "DataEmitterProcess": {
+          "text": "DataEmitterProcess"
+        },
+        "DataGeneration": {
+          "text": "DataGeneration"
+        },
+        "DissolvingProcess": {
+          "text": "DissolvingProcess"
+        },
+        "Extraction": {
+          "text": "Extraction"
+        },
+        "FiltrationProcess": {
+          "text": "FiltrationProcess"
+        },
+        "LibraryPreparation": {
+          "text": "LibraryPreparation"
+        },
+        "MagsAnalysis": {
+          "text": "MagsAnalysis"
+        },
+        "MassSpectrometry": {
+          "text": "MassSpectrometry"
+        },
+        "MaterialProcessing": {
+          "text": "MaterialProcessing"
+        },
+        "MetabolomicsAnalysis": {
+          "text": "MetabolomicsAnalysis"
+        },
+        "MetagenomeAnnotation": {
+          "text": "MetagenomeAnnotation"
+        },
+        "MetagenomeAssembly": {
+          "text": "MetagenomeAssembly"
+        },
+        "MetaproteomicsAnalysis": {
+          "text": "MetaproteomicsAnalysis"
+        },
+        "MetatranscriptomeAnnotation": {
+          "text": "MetatranscriptomeAnnotation"
+        },
+        "MetatranscriptomeAssembly": {
+          "text": "MetatranscriptomeAssembly"
+        },
+        "MetatranscriptomeExpressionAnalysis": {
+          "text": "MetatranscriptomeExpressionAnalysis"
+        },
+        "MixingProcess": {
+          "text": "MixingProcess"
+        },
+        "NomAnalysis": {
+          "text": "NomAnalysis"
+        },
+        "NucleotideSequencing": {
+          "text": "NucleotideSequencing"
+        },
+        "PlannedProcess": {
+          "text": "PlannedProcess"
+        },
+        "Pooling": {
+          "text": "Pooling"
+        },
+        "ReadBasedTaxonomyAnalysis": {
+          "text": "ReadBasedTaxonomyAnalysis"
+        },
+        "ReadQcAnalysis": {
+          "text": "ReadQcAnalysis"
+        },
+        "StorageProcess": {
+          "text": "StorageProcess"
+        },
+        "SubSamplingProcess": {
+          "text": "SubSamplingProcess"
+        },
+        "WorkflowExecution": {
+          "text": "WorkflowExecution"
+        }
+      }
+    },
     "SampleTypeEnum": {
       "name": "SampleTypeEnum",
       "from_schema": "https://w3id.org/nmdc/nmdc",
@@ -8747,7 +8847,7 @@
       "annotations": {
         "storage_units": {
           "tag": "storage_units",
-          "value": "%|mmol/L|umol/L"
+          "value": "%|mmol/L|umol/L|mg/L|g/L"
         }
       },
       "description": "When solutions A (containing substance X) and B are combined together, this slot captures the concentration of X in the combination",
@@ -9208,8 +9308,7 @@
         "qud:unit",
         "schema:unitCode"
       ],
-      "range": "UnitEnum",
-      "required": true
+      "range": "string"
     },
     "type": {
       "name": "type",
@@ -9335,6 +9434,53 @@
       "multivalued": true,
       "pattern": "^[Hh][Tt][Tt][Pp][Ss]?:\\/\\/(?!.*[Dd][Oo][Ii]\\.[Oo][Rr][Gg]).*$"
     },
+    "has_attribute_label": {
+      "name": "has_attribute_label",
+      "description": "Human-readable label for the property (e.g., MIxS label, ENVO term label).",
+      "comments": [
+        "This provides a human-friendly name for the asserted property. For example, \"bicarbonate ion concentration\" or \"total phosphorus\". The label helps with readability and data discovery."
+      ],
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "range": "string"
+    },
+    "has_attribute_id": {
+      "name": "has_attribute_id",
+      "description": "CURIE or IRI for the property (MIxS slot, NMDC slot, ENVO/OBI term, etc.).",
+      "comments": [
+        "This provides a resolvable identifier for the property being asserted. Examples include MIXS:0000117 for total phosphorus, or ENVO:01001357 for bicarbonate ion concentration.",
+        "Prefer using standard ontology terms (ENVO, PATO, OBI, etc.) or MIxS identifiers when available to enhance interoperability."
+      ],
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "range": "uriorcurie"
+    },
+    "has_quantity_kind_id": {
+      "name": "has_quantity_kind_id",
+      "description": "Optional CURIE or IRI for the physical quantity kind (e.g., qudt:QuantityKind).",
+      "comments": [
+        "This slot enables precise semantic description of what physical quantity is being measured, independent of the specific units used. For example, qudt:MassConcentration or qudt:Temperature.",
+        "Using quantity kind identifiers from QUDT or similar vocabularies improves data integration and enables automated unit conversion."
+      ],
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "range": "uriorcurie"
+    },
+    "has_value_term_id": {
+      "name": "has_value_term_id",
+      "description": "CURIE or IRI for categorical values (ENVO, PATO, METPO, etc.).",
+      "comments": [
+        "Use this slot when the value of the property is a controlled vocabulary term rather than a numeric or free-text value. For example, ENVO:00002297 for \"desert ecosystem\" or PATO:0001199 for \"dry\"."
+      ],
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "range": "uriorcurie"
+    },
+    "has_datetime_value": {
+      "name": "has_datetime_value",
+      "description": "Date-time value for the property in ISO-8601 format.",
+      "comments": [
+        "Use this slot for temporal properties. The value should follow ISO-8601 format (e.g., \"2025-06-12T14:30:00Z\")."
+      ],
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "range": "string"
+    },
     "processing_institution_workflow_metadata": {
       "name": "processing_institution_workflow_metadata",
       "description": "Information about how workflow results were generated when the processing is done by an external organziation (e.g., JGI) such as software tool name and version or pipeline name and version.",
@@ -17871,8 +18017,11 @@
           "value": "m"
         }
       },
-      "description": "Any other measurement performed or parameter collected, that is not listed here",
+      "description": "Structured miscellaneous property assertions for this Biosample. Use when a value cannot cleanly fit an existing, policy-governed slot.",
       "title": "miscellaneous parameter",
+      "todos": [
+        "This slot should not be available in the submission portal."
+      ],
       "examples": [
         {
           "value": "Bicarbonate ion concentration;2075 micromole per kilogram"
@@ -17885,7 +18034,7 @@
       "is_a": "core field",
       "string_serialization": "{text};{float} {unit}",
       "slot_uri": "MIXS:0000752",
-      "range": "TextValue",
+      "range": "PropertyAssertion",
       "multivalued": true,
       "inlined_as_list": true
     },
@@ -26084,6 +26233,12 @@
       "from_schema": "https://w3id.org/nmdc/nmdc",
       "range": "FailureWhatEnum"
     },
+    "protocol_for": {
+      "name": "protocol_for",
+      "description": "The type of planned process that the protocol describes.",
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "range": "ProtocolForEnum"
+    },
     "emsl_store_temp": {
       "name": "emsl_store_temp",
       "description": "The temperature at which the sample should be stored upon delivery to EMSL",
@@ -29515,7 +29670,9 @@
         },
         "has_unit": {
           "name": "has_unit",
-          "description": "The unit of the quantity"
+          "description": "The unit of the quantity",
+          "range": "UnitEnum",
+          "required": true
         },
         "has_numeric_value": {
           "name": "has_numeric_value",
@@ -29662,6 +29819,77 @@
       },
       "class_uri": "nmdc:GeolocationValue"
     },
+    "PropertyAssertion": {
+      "name": "PropertyAssertion",
+      "description": "A structured record of data that doesn't fit nicely within the constraints of other NMDC AttributeValues. Uses primitive ranges only.",
+      "comments": [
+        "This class enables flexible metadata capture for properties that don't align with existing, policy-governed slots in NMDC schema.",
+        "Interoperability note: This approach is aligned with the BERtron schema's 'properties' slot pattern (see https://github.com/ber-data/bertron-schema), which uses any_of to support both TextValue and QuantityValue ranges. NMDC's PropertyAssertion provides a more granular approach by using primitive types with optional semantic annotations, enabling better validation while maintaining flexibility for diverse metadata types.",
+        "PropertyAssertion supports both categorical values (via has_value_term_id) and numeric values (via has_numeric_value with optional min/max ranges), along with temporal (has_datetime_value) and boolean (has_boolean_value) data types."
+      ],
+      "examples": [
+        {
+          "description": "Data provided from submission that doesn't conform to required abs_air_humidity units",
+          "object": {
+            "type": "nmdc:PropertyAssertion",
+            "has_raw_value": "50 kPa",
+            "has_attribute_id": "MIXS:0000122",
+            "has_attribute_label": "absolute air humidity",
+            "has_numeric_value": 50,
+            "has_unit": "kPa"
+          }
+        },
+        {
+          "description": "Data provided from submission that doesn't conform to required UCUM unit",
+          "object": {
+            "type": "nmdc:PropertyAssertion",
+            "has_raw_value": "5.5 mL/L",
+            "has_attribute_id": "MIXS:0000119",
+            "has_attribute_label": "dissolved oxygen",
+            "has_numeric_value": 5.5,
+            "has_unit": "mL/L"
+          }
+        },
+        {
+          "description": "Unit provided is invalid for UCUM and standard",
+          "object": {
+            "type": "nmdc:PropertyAssertion",
+            "has_raw_value": "250 W/m2",
+            "has_attribute_id": "MIXS:0000112",
+            "has_attribute_label": "solar irradiance",
+            "has_numeric_value": 250,
+            "has_unit": "W/m2"
+          }
+        }
+      ],
+      "from_schema": "https://w3id.org/nmdc/nmdc",
+      "is_a": "AttributeValue",
+      "slots": [
+        "has_attribute_label",
+        "has_attribute_id",
+        "has_quantity_kind_id",
+        "has_value_term_id",
+        "has_boolean_value",
+        "has_datetime_value",
+        "has_numeric_value",
+        "has_minimum_numeric_value",
+        "has_maximum_numeric_value",
+        "has_unit"
+      ],
+      "slot_usage": {
+        "has_raw_value": {
+          "name": "has_raw_value",
+          "description": "Original contributor string representation (unparsed)",
+          "required": true
+        },
+        "has_unit": {
+          "name": "has_unit",
+          "description": "UCUM unit code (required only when numeric value is present)",
+          "required": false
+        }
+      },
+      "class_uri": "nmdc:PropertyAssertion"
+    },
     "NamedThing": {
       "name": "NamedThing",
       "description": "a databased entity or concept/class",
@@ -29763,7 +29991,7 @@
     },
     "Instrument": {
       "name": "Instrument",
-      "description": "A material entity that is designed to perform a function in a scientific investigation, but is not a reagent.",
+      "description": "A material entity that is designed to perform a function in a scientific investigation, but is not a reagent. This class models the make and model of the instrument, not the specific instance of the instrument.",
       "from_schema": "https://w3id.org/nmdc/nmdc",
       "aliases": [
         "device"
@@ -29786,7 +30014,17 @@
           }
         }
       },
-      "class_uri": "nmdc:Instrument"
+      "class_uri": "nmdc:Instrument",
+      "unique_keys": {
+        "main": {
+          "unique_key_name": "main",
+          "unique_key_slots": [
+            "vendor",
+            "model"
+          ],
+          "description": "A unique instrument is defined by its vendor and model."
+        }
+      }
     },
     "PlannedProcess": {
       "name": "PlannedProcess",
@@ -29814,7 +30052,9 @@
         "url",
         "name",
         "type",
-        "description"
+        "description",
+        "protocol_for",
+        "analysis_type"
       ],
       "class_uri": "nmdc:Protocol"
     },