hpo¶
knowledge_graph.clinical.hpo
HPO ontology, and the comparison of three HPO linkers on the SYMPTOM mentions of the ten test notes.
python -m knowledge_graph.clinical.hpo prepare # mentions to link and the gold links python -m knowledge_graph.clinical.hpo fasthpocr # FastHPOCR longest match inside each mention python -m knowledge_graph.clinical.hpo phenobert # PhenoBERT dictionary, then its fine-tuned phrase matcher python -m knowledge_graph.clinical.hpo sapbert # SapBERT nearest alias, threshold chosen on validation
normalize_surface(value: str) -> str
¶
Lowercase, with punctuation and underscores turned into single spaces.
Source code in src/knowledge_graph/clinical/hpo.py
23 24 25 | |
Term
dataclass
¶
HpoOntology
¶
The terms of the HPO OBO file.
Source code in src/knowledge_graph/clinical/hpo.py
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 | |
path = path
instance-attribute
¶
version = 'unknown'
instance-attribute
¶
names = {}
instance-attribute
¶
parents = {}
instance-attribute
¶
terms = {}
instance-attribute
¶
alternate = {}
instance-attribute
¶
canonical_id(value: str | None) -> str | None
¶
The current term of an ID or alternate ID, or None.
Source code in src/knowledge_graph/clinical/hpo.py
68 69 70 71 | |
ancestors(term_id: str) -> set[str]
¶
Source code in src/knowledge_graph/clinical/hpo.py
73 74 75 76 77 78 79 80 | |
gold_links(ontology, ids)
¶
Reference HPO links of SYMPTOM mentions: (note, begin, end, HPO ID).
Source code in src/knowledge_graph/clinical/hpo.py
83 84 85 86 87 88 89 | |
mention_record(note_id, text, begin, end)
¶
Source code in src/knowledge_graph/clinical/hpo.py
92 93 | |
unique_spans(mentions)
¶
One mention per span (the last one wins).
Source code in src/knowledge_graph/clinical/hpo.py
96 97 98 | |
prepare(out)
¶
Mentions to link: all reference SYMPTOM mentions, and the symptom predictions of GLiNER-Relex and HunFlair2.
Source code in src/knowledge_graph/clinical/hpo.py
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 | |
score(out, name, cache, threshold=-1)
¶
Test metrics and predicted links of every detector and input condition, from the phrase links in cache.
Source code in src/knowledge_graph/clinical/hpo.py
129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 | |
fasthpocr(out)
¶
Source code in src/knowledge_graph/clinical/hpo.py
146 147 148 149 150 151 152 153 154 155 156 157 158 | |
phenobert(out)
¶
PhenoBERT's exact dictionary match first, then its fine-tuned phrase matcher.
Source code in src/knowledge_graph/clinical/hpo.py
161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 | |
sapbert(out)
¶
SapBERT's nearest HPO alias, with the score threshold chosen on the validation reference mentions.
Source code in src/knowledge_graph/clinical/hpo.py
192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 | |
main()
¶
Source code in src/knowledge_graph/clinical/hpo.py
231 232 233 234 235 236 | |