Working with Complex Hierarchical Data¶

This tutorial demonstrates how linkml-data-qc handles deeply nested data structures with multiple levels of lists and objects. We'll use a disease knowledge base as an example.

The Challenge¶

Real-world knowledge bases often have complex structures:

Diseases with multiple subtypes
Each subtype has evidence citations
Pathophysiology entries with cell type annotations
Phenotypes with nested ontology term bindings

Tracking compliance across all these nested structures requires understanding path notation.

Setup: A Complex Schema¶

Let's create a schema that models a disease knowledge base with nested structures:

In [1]:

Copied!





%%bash
cat > /tmp/disease_kb_schema.yaml << 'EOF'
id: https://example.org/disease-kb
name: disease_kb_schema
prefixes:
  linkml: https://w3id.org/linkml/
imports:
  - linkml:types
default_range: string

classes:
  Disease:
    attributes:
      name:
        required: true
      description:
        recommended: true
      disease_term:
        range: OntologyAnnotation
        inlined: true
        recommended: true
      has_subtypes:
        range: DiseaseSubtype
        multivalued: true
        inlined_as_list: true
      pathophysiology:
        range: PathophysiologyStep
        multivalued: true
        inlined_as_list: true
      phenotypes:
        range: Phenotype
        multivalued: true
        inlined_as_list: true

  OntologyAnnotation:
    attributes:
      preferred_term:
        recommended: true
      term:
        range: OntologyTerm
        inlined: true
        recommended: true

  OntologyTerm:
    attributes:
      id:
        recommended: true
      label:
        recommended: true

  DiseaseSubtype:
    attributes:
      name:
        required: true
      description:
        recommended: true
      evidence:
        range: Evidence
        multivalued: true
        inlined_as_list: true

  Evidence:
    attributes:
      reference:
        recommended: true
      supports:
        recommended: true
      snippet:
        recommended: true
      explanation:
        recommended: true

  PathophysiologyStep:
    attributes:
      name:
        required: true
      description:
        recommended: true
      cell_types:
        range: OntologyAnnotation
        multivalued: true
        inlined_as_list: true
      evidence:
        range: Evidence
        multivalued: true
        inlined_as_list: true

  Phenotype:
    attributes:
      name:
        required: true
      category:
        recommended: true
      frequency:
        recommended: true
      phenotype_term:
        range: OntologyAnnotation
        inlined: true
        recommended: true
      evidence:
        range: Evidence
        multivalued: true
        inlined_as_list: true
EOF
echo "Schema created with nested classes!"
%%bash
cat > /tmp/disease_kb_schema.yaml << 'EOF'
id: https://example.org/disease-kb
name: disease_kb_schema
prefixes:
  linkml: https://w3id.org/linkml/
imports:
  - linkml:types
default_range: string

classes:
  Disease:
    attributes:
      name:
        required: true
      description:
        recommended: true
      disease_term:
        range: OntologyAnnotation
        inlined: true
        recommended: true
      has_subtypes:
        range: DiseaseSubtype
        multivalued: true
        inlined_as_list: true
      pathophysiology:
        range: PathophysiologyStep
        multivalued: true
        inlined_as_list: true
      phenotypes:
        range: Phenotype
        multivalued: true
        inlined_as_list: true

  OntologyAnnotation:
    attributes:
      preferred_term:
        recommended: true
      term:
        range: OntologyTerm
        inlined: true
        recommended: true

  OntologyTerm:
    attributes:
      id:
        recommended: true
      label:
        recommended: true

  DiseaseSubtype:
    attributes:
      name:
        required: true
      description:
        recommended: true
      evidence:
        range: Evidence
        multivalued: true
        inlined_as_list: true

  Evidence:
    attributes:
      reference:
        recommended: true
      supports:
        recommended: true
      snippet:
        recommended: true
      explanation:
        recommended: true

  PathophysiologyStep:
    attributes:
      name:
        required: true
      description:
        recommended: true
      cell_types:
        range: OntologyAnnotation
        multivalued: true
        inlined_as_list: true
      evidence:
        range: Evidence
        multivalued: true
        inlined_as_list: true

  Phenotype:
    attributes:
      name:
        required: true
      category:
        recommended: true
      frequency:
        recommended: true
      phenotype_term:
        range: OntologyAnnotation
        inlined: true
        recommended: true
      evidence:
        range: Evidence
        multivalued: true
        inlined_as_list: true
EOF
echo "Schema created with nested classes!"

Schema created with nested classes!

Test Data: A Disease Entry¶

Now let's create a disease entry with varying levels of completeness:

In [2]:

Copied!





%%bash
cat > /tmp/disease_aps.yaml << 'EOF'
name: Antiphospholipid Syndrome
description: A systemic autoimmune disorder characterized by antiphospholipid antibodies.
disease_term:
  preferred_term: antiphospholipid syndrome
  term:
    id: MONDO:0000810
    label: antiphospholipid syndrome

has_subtypes:
  - name: Primary APS
    description: occurs in the absence of any other disease
    evidence:
      - reference: PMID:16338214
        supports: SUPPORT
        snippet: the condition can exist on its own
        explanation: Supports primary APS definition
      - reference: PMID:27550302
        supports: SUPPORT
        snippet: APS can be isolated (primary APS)
        # Missing explanation!
  - name: Secondary APS
    description: occurs with other autoimmune diseases
    evidence:
      - reference: PMID:11014973
        supports: SUPPORT
        snippet: APS may be associated with another autoimmune disease
        explanation: Confirms secondary APS
  - name: Asymptomatic APS
    # Missing description!
    evidence:
      - reference: PMID:17145604
        supports: SUPPORT
        # Missing snippet and explanation!

pathophysiology:
  - name: Antibody Production
    description: The immune system produces antiphospholipid antibodies.
    cell_types:
      - preferred_term: B cell
        term:
          id: CL:0000236
          label: B cell
      - preferred_term: T cell
        # Missing nested term!
    evidence:
      - reference: PMID:29867951
        supports: SUPPORT
        snippet: production of antibodies that bind phospholipid-binding proteins
        explanation: Describes antibody production mechanism
  - name: Blood Clot Formation
    description: Antibodies increase thrombosis risk.
    # No cell_types - that's fine, not always applicable
    evidence:
      - reference: PMID:22100379
        supports: SUPPORT
        snippet: antibodies associated with thrombosis risk
        explanation: Supports clotting mechanism

phenotypes:
  - name: Deep Vein Thrombosis
    category: Thrombosis
    frequency: FREQUENT
    phenotype_term:
      preferred_term: DVT
      term:
        id: HP:0002625
        label: deep vein thrombosis
    evidence:
      - reference: PMID:36575066
        supports: SUPPORT
        snippet: venous thromboembolism is frequent
        explanation: Confirms DVT frequency
  - name: Preeclampsia
    category: Pregnancy-Related
    # Missing frequency!
    phenotype_term:
      preferred_term: Preeclampsia
      term:
        id: HP:0100602
        # Missing label!
    evidence:
      - reference: PMID:26815583
        supports: PARTIAL
        # Missing snippet and explanation!
  - name: Livedo Reticularis
    category: Dermatologic
    frequency: OCCASIONAL
    # Missing phenotype_term entirely!
    evidence:
      - reference: PMID:26223086
        supports: SUPPORT
        snippet: common cutaneous manifestation
        explanation: Supports dermatologic involvement
EOF
echo "Complex disease entry created!"
%%bash
cat > /tmp/disease_aps.yaml << 'EOF'
name: Antiphospholipid Syndrome
description: A systemic autoimmune disorder characterized by antiphospholipid antibodies.
disease_term:
  preferred_term: antiphospholipid syndrome
  term:
    id: MONDO:0000810
    label: antiphospholipid syndrome

has_subtypes:
  - name: Primary APS
    description: occurs in the absence of any other disease
    evidence:
      - reference: PMID:16338214
        supports: SUPPORT
        snippet: the condition can exist on its own
        explanation: Supports primary APS definition
      - reference: PMID:27550302
        supports: SUPPORT
        snippet: APS can be isolated (primary APS)
        # Missing explanation!
  - name: Secondary APS
    description: occurs with other autoimmune diseases
    evidence:
      - reference: PMID:11014973
        supports: SUPPORT
        snippet: APS may be associated with another autoimmune disease
        explanation: Confirms secondary APS
  - name: Asymptomatic APS
    # Missing description!
    evidence:
      - reference: PMID:17145604
        supports: SUPPORT
        # Missing snippet and explanation!

pathophysiology:
  - name: Antibody Production
    description: The immune system produces antiphospholipid antibodies.
    cell_types:
      - preferred_term: B cell
        term:
          id: CL:0000236
          label: B cell
      - preferred_term: T cell
        # Missing nested term!
    evidence:
      - reference: PMID:29867951
        supports: SUPPORT
        snippet: production of antibodies that bind phospholipid-binding proteins
        explanation: Describes antibody production mechanism
  - name: Blood Clot Formation
    description: Antibodies increase thrombosis risk.
    # No cell_types - that's fine, not always applicable
    evidence:
      - reference: PMID:22100379
        supports: SUPPORT
        snippet: antibodies associated with thrombosis risk
        explanation: Supports clotting mechanism

phenotypes:
  - name: Deep Vein Thrombosis
    category: Thrombosis
    frequency: FREQUENT
    phenotype_term:
      preferred_term: DVT
      term:
        id: HP:0002625
        label: deep vein thrombosis
    evidence:
      - reference: PMID:36575066
        supports: SUPPORT
        snippet: venous thromboembolism is frequent
        explanation: Confirms DVT frequency
  - name: Preeclampsia
    category: Pregnancy-Related
    # Missing frequency!
    phenotype_term:
      preferred_term: Preeclampsia
      term:
        id: HP:0100602
        # Missing label!
    evidence:
      - reference: PMID:26815583
        supports: PARTIAL
        # Missing snippet and explanation!
  - name: Livedo Reticularis
    category: Dermatologic
    frequency: OCCASIONAL
    # Missing phenotype_term entirely!
    evidence:
      - reference: PMID:26223086
        supports: SUPPORT
        snippet: common cutaneous manifestation
        explanation: Supports dermatologic involvement
EOF
echo "Complex disease entry created!"

Complex disease entry created!

Analyzing Nested Compliance¶

Let's run the analysis and see how paths are reported:

In [3]:

Copied!





%%bash
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease
%%bash
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease

Compliance Report: /tmp/disease_aps.yaml
Target Class: Disease
Global Compliance: 85.7% (24/28)
Weig

hted Compliance: 82.6%

Summary by Slot:
  category: 100.0%
  disease_term: 100.0%
  frequency: 66.7

%
  id: 100.0%
  label: 75.0%
  phenotype_term: 66.7%
  preferred_term: 100.0%
  term: 80.0%

Aggreg

ated Scores by List Path:
  pathophysiology[].cell_types[].preferred_term: 100.0% (2/2)
  pathophysi

ology[].cell_types[].term: 50.0% (1/2)
  pathophysiology[].cell_types[].term.id: 100.0% (1/1)
  path

ophysiology[].cell_types[].term.label: 100.0% (1/1)
  phenotypes[].category: 100.0% (3/3)
  phenotyp

es[].frequency: 66.7% (2/3)
  phenotypes[].phenotype_term: 66.7% (2/3)
  phenotypes[].phenotype_term

.preferred_term: 100.0% (2/2)
  phenotypes[].phenotype_term.term: 100.0% (2/2)
  phenotypes[].phenot

ype_term.term.id: 100.0% (2/2)
  phenotypes[].phenotype_term.term.label: 50.0% (1/2)

Detailed Path

Scores:
  (root) (Disease): 100.0%
    - disease_term: OK
  disease_term (OntologyAnnotation): 100.0

%
    - preferred_term: OK
    - term: OK
  disease_term.term (OntologyTerm): 100.0%
    - id: OK

  - label: OK
  pathophysiology[0].cell_types[0] (OntologyAnnotation): 100.0%
    - preferred_term:

OK
    - term: OK
  pathophysiology[0].cell_types[0].term (OntologyTerm): 100.0%
    - id: OK
    -

label: OK
  pathophysiology[0].cell_types[1] (OntologyAnnotation): 50.0%
    - preferred_term: OK

  - term: MISSING
  phenotypes[0] (Phenotype): 100.0%
    - category: OK
    - frequency: OK
    - p

henotype_term: OK
  phenotypes[0].phenotype_term (OntologyAnnotation): 100.0%
    - preferred_term:

OK
    - term: OK
  phenotypes[0].phenotype_term.term (OntologyTerm): 100.0%
    - id: OK
    - labe

l: OK
  phenotypes[1] (Phenotype): 66.7%
    - category: OK
    - frequency: MISSING
    - phenotype

_term: OK
  phenotypes[1].phenotype_term (OntologyAnnotation): 100.0%
    - preferred_term: OK
    -

 term: OK
  phenotypes[1].phenotype_term.term (OntologyTerm): 50.0%
    - id: OK
    - label: MISSIN

G
  phenotypes[2] (Phenotype): 66.7%
    - category: OK
    - frequency: OK
    - phenotype_term: MI

SSING

Understanding the Paths¶

The output shows aggregated paths like:

has_subtypes[].description - All subtype descriptions
has_subtypes[].evidence[].snippet - All snippets across all evidence in all subtypes
pathophysiology[].cell_types[].term - All nested term objects in cell_types
phenotypes[].phenotype_term.term.label - Deeply nested labels

Let's look at the JSON output for more detail:

In [4]:

Copied!





%%bash
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease \
    -f json | python -m json.tool | head -100
%%bash
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease \
    -f json | python -m json.tool | head -100

{
    "file_path": "/tmp/disease_aps.yaml",
    "target_class": "Disease",
    "schema_path": "/tmp/

disease_kb_schema.yaml",
    "global_compliance": 85.71428571428571,
    "weighted_compliance": 82.6

086956521739,
    "total_checks": 28,
    "total_populated": 24,
    "path_scores": [
        {

        "path": "(root)",
            "parent_class": "Disease",
            "item_count": 1,

      "slot_scores": [
                {
                    "path": "(root)",
                    "

slot_name": "disease_term",
                    "populated": 1,
                    "total": 1,

                "percentage": 100.0
                }
            ],
            "overall_percentage

": 100.0
        },
        {
            "path": "disease_term",
            "parent_class": "Ontol

ogyAnnotation",
            "item_count": 1,
            "slot_scores": [
                {

            "path": "disease_term",
                    "slot_name": "preferred_term",

       "populated": 1,
                    "total": 1,
                    "percentage": 100.0

           },
                {
                    "path": "disease_term",
                    "slo

t_name": "term",
                    "populated": 1,
                    "total": 1,

     "percentage": 100.0
                }
            ],
            "overall_percentage": 100.0

      },
        {
            "path": "disease_term.term",
            "parent_class": "OntologyTer

m",
            "item_count": 1,
            "slot_scores": [
                {

"path": "disease_term.term",
                    "slot_name": "id",
                    "populated":

 1,
                    "total": 1,
                    "percentage": 100.0
                },

           {
                    "path": "disease_term.term",
                    "slot_name": "labe

l",
                    "populated": 1,
                    "total": 1,
                    "percent

age": 100.0
                }
            ],
            "overall_percentage": 100.0
        },

    {
            "path": "pathophysiology[0].cell_types[0]",
            "parent_class": "OntologyA

nnotation",
            "item_count": 1,
            "slot_scores": [
                {

        "path": "pathophysiology[0].cell_types[0]",
                    "slot_name": "preferred_term

",
                    "populated": 1,
                    "total": 1,
                    "percenta

ge": 100.0
                },
                {
                    "path": "pathophysiology[0].cell

_types[0]",
                    "slot_name": "term",
                    "populated": 1,

         "total": 1,
                    "percentage": 100.0
                }
            ],

      "overall_percentage": 100.0
        },
        {
            "path": "pathophysiology[0].cell_

types[0].term",
            "parent_class": "OntologyTerm",
            "item_count": 1,

 "slot_scores": [
                {
                    "path": "pathophysiology[0].cell_types[0].te

rm",
                    "slot_name": "id",
                    "populated": 1,

"total": 1,

Configuring Weights for Nested Paths¶

We can prioritize certain nested paths over others:

In [5]:

Copied!





%%bash
cat > /tmp/disease_qc_config.yaml << 'EOF'
default_weight: 1.0

# Slot-level configuration
slots:
  # Ontology term bindings are critical
  id:
    weight: 3.0
    min_compliance: 90.0
  label:
    weight: 2.0
    min_compliance: 80.0
  # Evidence snippets are important
  snippet:
    weight: 2.0
    min_compliance: 75.0
  explanation:
    weight: 1.5
  # Descriptions nice-to-have
  description:
    weight: 0.5

# Path-level overrides for specific nested locations
paths:
  # Phenotype terms are especially critical
  "phenotypes[].phenotype_term.term.id":
    weight: 5.0
    min_compliance: 95.0
  # Root disease term is essential
  "disease_term.term.id":
    weight: 5.0
    min_compliance: 100.0
EOF
echo "Configuration created!"
%%bash
cat > /tmp/disease_qc_config.yaml << 'EOF'
default_weight: 1.0

# Slot-level configuration
slots:
  # Ontology term bindings are critical
  id:
    weight: 3.0
    min_compliance: 90.0
  label:
    weight: 2.0
    min_compliance: 80.0
  # Evidence snippets are important
  snippet:
    weight: 2.0
    min_compliance: 75.0
  explanation:
    weight: 1.5
  # Descriptions nice-to-have
  description:
    weight: 0.5

# Path-level overrides for specific nested locations
paths:
  # Phenotype terms are especially critical
  "phenotypes[].phenotype_term.term.id":
    weight: 5.0
    min_compliance: 95.0
  # Root disease term is essential
  "disease_term.term.id":
    weight: 5.0
    min_compliance: 100.0
EOF
echo "Configuration created!"

Configuration created!

In [6]:

Copied!





%%bash
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease \
    -c /tmp/disease_qc_config.yaml
%%bash
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease \
    -c /tmp/disease_qc_config.yaml

Compliance Report: /tmp/disease_aps.yaml
Target Class: Disease
Global Compliance: 85.7% (24/28)
Weig

hted Compliance: 86.1%
Config: /tmp/disease_qc_config.yaml

Summary by Slot:
  category: 100.0%
  di

sease_term: 100.0%
  frequency: 66.7%
  id: 100.0%
  label: 75.0%
  phenotype_term: 66.7%
  preferre

d_term: 100.0%
  term: 80.0%

Threshold Violations (1):
  phenotypes[].phenotype_term.term.label: 50

.0% < 80.0% (shortfall: 30.0%)

Aggregated Scores by List Path:
  pathophysiology[].cell_types[].pre

ferred_term: 100.0% (2/2)
  pathophysiology[].cell_types[].term: 50.0% (1/2)
  pathophysiology[].cel

l_types[].term.id: 100.0% (1/1)
  pathophysiology[].cell_types[].term.label: 100.0% (1/1)
  phenotyp

es[].category: 100.0% (3/3)
  phenotypes[].frequency: 66.7% (2/3)
  phenotypes[].phenotype_term: 66.

7% (2/3)
  phenotypes[].phenotype_term.preferred_term: 100.0% (2/2)
  phenotypes[].phenotype_term.te

rm: 100.0% (2/2)
  phenotypes[].phenotype_term.term.id: 100.0% (2/2)
  phenotypes[].phenotype_term.t

erm.label: 50.0% (1/2)

Detailed Path Scores:
  (root) (Disease): 100.0%
    - disease_term: OK
  di

sease_term (OntologyAnnotation): 100.0%
    - preferred_term: OK
    - term: OK
  disease_term.term

(OntologyTerm): 100.0%
    - id: OK
    - label: OK
  pathophysiology[0].cell_types[0] (OntologyAnno

tation): 100.0%
    - preferred_term: OK
    - term: OK
  pathophysiology[0].cell_types[0].term (Ont

ologyTerm): 100.0%
    - id: OK
    - label: OK
  pathophysiology[0].cell_types[1] (OntologyAnnotati

on): 50.0%
    - preferred_term: OK
    - term: MISSING
  phenotypes[0] (Phenotype): 100.0%
    - ca

tegory: OK
    - frequency: OK
    - phenotype_term: OK
  phenotypes[0].phenotype_term (OntologyAnno

tation): 100.0%
    - preferred_term: OK
    - term: OK
  phenotypes[0].phenotype_term.term (Ontolog

yTerm): 100.0%
    - id: OK
    - label: OK
  phenotypes[1] (Phenotype): 66.7%
    - category: OK

  - frequency: MISSING
    - phenotype_term: OK
  phenotypes[1].phenotype_term (OntologyAnnotation):

 100.0%
    - preferred_term: OK
    - term: OK
  phenotypes[1].phenotype_term.term (OntologyTerm):

50.0%
    - id: OK
    - label: MISSING
  phenotypes[2] (Phenotype): 66.7%
    - category: OK
    -

frequency: OK
    - phenotype_term: MISSING

Checking for Threshold Violations¶

In [7]:

Copied!





%%bash
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease \
    -c /tmp/disease_qc_config.yaml \
    --fail-on-violations || echo "Violations detected! Exit code: $?"
%%bash
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease \
    -c /tmp/disease_qc_config.yaml \
    --fail-on-violations || echo "Violations detected! Exit code: $?"

Compliance Report: /tmp/disease_aps.yaml
Target Class: Disease
Global Compliance: 85.7% (24/28)
Weig

hted Compliance: 86.1%
Config: /tmp/disease_qc_config.yaml

Summary by Slot:
  category: 100.0%
  di

sease_term: 100.0%
  frequency: 66.7%
  id: 100.0%
  label: 75.0%
  phenotype_term: 66.7%
  preferre

d_term: 100.0%
  term: 80.0%

Threshold Violations (1):
  phenotypes[].phenotype_term.term.label: 50

.0% < 80.0% (shortfall: 30.0%)

Aggregated Scores by List Path:
  pathophysiology[].cell_types[].pre

ferred_term: 100.0% (2/2)
  pathophysiology[].cell_types[].term: 50.0% (1/2)
  pathophysiology[].cel

l_types[].term.id: 100.0% (1/1)
  pathophysiology[].cell_types[].term.label: 100.0% (1/1)
  phenotyp

es[].category: 100.0% (3/3)
  phenotypes[].frequency: 66.7% (2/3)
  phenotypes[].phenotype_term: 66.

7% (2/3)
  phenotypes[].phenotype_term.preferred_term: 100.0% (2/2)
  phenotypes[].phenotype_term.te

rm: 100.0% (2/2)
  phenotypes[].phenotype_term.term.id: 100.0% (2/2)
  phenotypes[].phenotype_term.t

erm.label: 50.0% (1/2)

Detailed Path Scores:
  (root) (Disease): 100.0%
    - disease_term: OK
  di

sease_term (OntologyAnnotation): 100.0%
    - preferred_term: OK
    - term: OK
  disease_term.term

(OntologyTerm): 100.0%
    - id: OK
    - label: OK
  pathophysiology[0].cell_types[0] (OntologyAnno

tation): 100.0%
    - preferred_term: OK
    - term: OK
  pathophysiology[0].cell_types[0].term (Ont

ologyTerm): 100.0%
    - id: OK
    - label: OK
  pathophysiology[0].cell_types[1] (OntologyAnnotati

on): 50.0%
    - preferred_term: OK
    - term: MISSING
  phenotypes[0] (Phenotype): 100.0%
    - ca

tegory: OK
    - frequency: OK
    - phenotype_term: OK
  phenotypes[0].phenotype_term (OntologyAnno

tation): 100.0%
    - preferred_term: OK
    - term: OK
  phenotypes[0].phenotype_term.term (Ontolog

yTerm): 100.0%
    - id: OK
    - label: OK
  phenotypes[1] (Phenotype): 66.7%
    - category: OK

  - frequency: MISSING
    - phenotype_term: OK
  phenotypes[1].phenotype_term (OntologyAnnotation):

 100.0%
    - preferred_term: OK
    - term: OK
  phenotypes[1].phenotype_term.term (OntologyTerm):

50.0%
    - id: OK
    - label: MISSING
  phenotypes[2] (Phenotype): 66.7%
    - category: OK
    -

frequency: OK
    - phenotype_term: MISSING

1 threshold violation(s) found:
  phenotypes[].phenotype_term.term.label: 50.0% < 80.0%

Violations detected! Exit code: 1

Finding Low-Compliance Areas¶

Use CSV output to find the paths with lowest compliance:

In [8]:

Copied!





%%bash
echo "=== Lowest compliance areas ==="
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease \
    -f csv | tail -n +2 | sort -t, -k7 -n | head -10
%%bash
echo "=== Lowest compliance areas ==="
linkml-data-qc /tmp/disease_aps.yaml \
    -s /tmp/disease_kb_schema.yaml \
    -t Disease \
    -f csv | tail -n +2 | sort -t, -k7 -n | head -10

=== Lowest compliance areas ===

/tmp/disease_aps.yaml,pathophysiology[0].cell_types[1],OntologyAnnotation,term,0,1,0.0
/tmp/diseas

e_aps.yaml,phenotypes[1],Phenotype,frequency,0,1,0.0
/tmp/disease_aps.yaml,phenotypes[1].phenotype_

term.term,OntologyTerm,label,0,1,0.0
/tmp/disease_aps.yaml,phenotypes[2],Phenotype,phenotype_term,0

,1,0.0
/tmp/disease_aps.yaml,(root),Disease,disease_term,1,1,100.0
/tmp/disease_aps.yaml,disease_t

erm,OntologyAnnotation,preferred_term,1,1,100.0
/tmp/disease_aps.yaml,disease_term,OntologyAnnotati

on,term,1,1,100.0
/tmp/disease_aps.yaml,disease_term.term,OntologyTerm,id,1,1,100.0
/tmp/disease_a

ps.yaml,disease_term.term,OntologyTerm,label,1,1,100.0

Path Patterns Summary¶

Here's what we learned about paths in complex data:

Pattern	Example	Meaning
`(root)`	-	Top-level object
`slot`	`disease_term`	Direct child object
`slot.nested`	`disease_term.term`	Nested object
`list[]`	`has_subtypes[]`	All items in list
`list[].slot`	`has_subtypes[].description`	Slot in all list items
`list[].nested[]`	`has_subtypes[].evidence[]`	Nested list in list
`list[].obj.slot`	`phenotypes[].phenotype_term.term`	Deeply nested

Key Takeaways¶

Aggregated paths (with []) show compliance across all matching items
Deeply nested paths help identify specific curation gaps
Path-specific config lets you prioritize critical nested locations
CSV output is useful for finding lowest-compliance areas to prioritize curation