{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "c91970a3",
   "metadata": {},
   "source": [
    "# Ontology Term Embeddings in OAK\n",
    "\n",
    "This notebook introduces the `EmbeddingProviderInterface`, which gives direct access to\n",
    "**vector embeddings of ontology terms** as numpy matrices and pandas DataFrames, along\n",
    "with operations built on them (similarity, nearest neighbours, text search, and\n",
    "best-match comparison of term sets).\n",
    "\n",
    "Backends:\n",
    "\n",
    "| Adapter | Models | Notes |\n",
    "|---|---|---|\n",
    "| `ols:<ontology>` | precomputed LLM embeddings hosted by [OLS](https://www.ebi.ac.uk/ols4) | several models, 512-d (PCA-reduced); cached locally |\n",
    "| `llm:<adapter>` | any embedding model supported by the [llm](https://llm.datasette.io) library | embeds labels (or labels + definitions) on demand; cached locally |\n",
    "| any adapter that can compute ancestors (e.g. `sqlite:obo:hp`) | `closure` | each term is a multi-hot vector of its reflexive ancestors |\n",
    "\n",
    "The `closure` model is how \"classic\" ontology semantic similarity looks when written\n",
    "as vectors. Jaccard similarity of closure vectors **is** the familiar ancestor-set Jaccard,\n",
    "so classic and learned embeddings can be compared with exactly the same code."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "bf254b1f",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:13.990354Z",
     "iopub.status.busy": "2026-10-05T01:10:13.989953Z",
     "iopub.status.idle": "2026-10-05T01:10:15.726047Z",
     "shell.execute_reply": "2026-10-05T01:10:15.724453Z"
    }
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\", category=UserWarning, module=\"eutils\")\n",
    "warnings.filterwarnings(\"ignore\", category=DeprecationWarning)\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "# reference palette (categorical slots in fixed order; sequential blue ramp)\n",
    "BLUE, ORANGE, AQUA = \"#2a78d6\", \"#eb6834\", \"#1baf7a\"\n",
    "INK, INK2, GRID, SURFACE = \"#0b0b0b\", \"#52514e\", \"#e4e3df\", \"#fcfcfb\"\n",
    "plt.rcParams.update({\n",
    "    \"figure.facecolor\": SURFACE, \"axes.facecolor\": SURFACE, \"savefig.facecolor\": SURFACE,\n",
    "    \"axes.edgecolor\": GRID, \"axes.labelcolor\": INK2, \"axes.titlecolor\": INK,\n",
    "    \"axes.grid\": True, \"grid.color\": GRID, \"grid.linewidth\": 0.6,\n",
    "    \"axes.spines.top\": False, \"axes.spines.right\": False,\n",
    "    \"xtick.color\": INK2, \"ytick.color\": INK2, \"text.color\": INK,\n",
    "    \"font.size\": 10, \"axes.titlesize\": 11, \"figure.dpi\": 110,\n",
    "})\n",
    "SEQUENTIAL = sns.blend_palette([\"#f0efec\", \"#86b6ef\", \"#2a78d6\", \"#0d366b\"], as_cmap=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7080aa07",
   "metadata": {},
   "source": [
    "## Connecting to OLS\n",
    "\n",
    "OLS serves embeddings for every class it indexes, for several models:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "846ab080",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:15.730651Z",
     "iopub.status.busy": "2026-10-05T01:10:15.730193Z",
     "iopub.status.idle": "2026-10-05T01:10:19.502703Z",
     "shell.execute_reply": "2026-10-05T01:10:19.501113Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['harrier-oss-v1-27b_pca512',\n",
       " 'llama-embed-nemotron-8b_pca512',\n",
       " 'text-embedding-3-large_pca512',\n",
       " 'text-embedding-3-small_pca512']"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from oaklib import get_adapter\n",
    "\n",
    "ols = get_adapter(\"ols:hp\")\n",
    "ols.embedding_models()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0a69374d",
   "metadata": {},
   "source": [
    "## Vectors as numpy and pandas\n",
    "\n",
    "`entity_embeddings` returns the CURIEs that have vectors and a matrix with one row per\n",
    "CURIE. Vectors are cached in a local sqlite database (under `~/.data/oaklib/embeddings`),\n",
    "so OLS is only asked once per term and model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "6f4c2fe6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:19.505970Z",
     "iopub.status.busy": "2026-10-05T01:10:19.505411Z",
     "iopub.status.idle": "2026-10-05T01:10:19.515680Z",
     "shell.execute_reply": "2026-10-05T01:10:19.514179Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10, 512)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hand_terms = [\n",
    "    \"HP:0001159\",  # Syndactyly\n",
    "    \"HP:0006101\",  # Finger syndactyly\n",
    "    \"HP:0001770\",  # Toe syndactyly\n",
    "    \"HP:0001161\",  # Hand polydactyly\n",
    "    \"HP:0001829\",  # Foot polydactyly\n",
    "    \"HP:0001155\",  # Abnormality of the hand\n",
    "    \"HP:0001760\",  # Abnormal foot morphology\n",
    "    \"HP:0000568\",  # Microphthalmia\n",
    "    \"HP:0000478\",  # Abnormality of the eye\n",
    "    \"HP:0001250\",  # Seizure\n",
    "]\n",
    "ids, matrix = ols.entity_embeddings(hand_terms)\n",
    "matrix.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "bd5211af",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:19.518637Z",
     "iopub.status.busy": "2026-10-05T01:10:19.518344Z",
     "iopub.status.idle": "2026-10-05T01:10:19.534211Z",
     "shell.execute_reply": "2026-10-05T01:10:19.532927Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "      <th>2</th>\n",
       "      <th>3</th>\n",
       "      <th>4</th>\n",
       "      <th>5</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>HP:0001159</th>\n",
       "      <td>0.096</td>\n",
       "      <td>-0.007</td>\n",
       "      <td>-0.201</td>\n",
       "      <td>-0.119</td>\n",
       "      <td>-0.137</td>\n",
       "      <td>0.003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP:0006101</th>\n",
       "      <td>0.064</td>\n",
       "      <td>-0.042</td>\n",
       "      <td>-0.216</td>\n",
       "      <td>-0.085</td>\n",
       "      <td>-0.159</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP:0001770</th>\n",
       "      <td>0.091</td>\n",
       "      <td>0.001</td>\n",
       "      <td>-0.189</td>\n",
       "      <td>-0.122</td>\n",
       "      <td>-0.142</td>\n",
       "      <td>-0.009</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP:0001161</th>\n",
       "      <td>0.068</td>\n",
       "      <td>-0.057</td>\n",
       "      <td>-0.183</td>\n",
       "      <td>-0.096</td>\n",
       "      <td>-0.152</td>\n",
       "      <td>-0.024</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP:0001829</th>\n",
       "      <td>0.084</td>\n",
       "      <td>-0.004</td>\n",
       "      <td>-0.178</td>\n",
       "      <td>-0.112</td>\n",
       "      <td>-0.172</td>\n",
       "      <td>-0.007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP:0001155</th>\n",
       "      <td>0.108</td>\n",
       "      <td>-0.077</td>\n",
       "      <td>-0.241</td>\n",
       "      <td>-0.038</td>\n",
       "      <td>-0.166</td>\n",
       "      <td>-0.016</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP:0001760</th>\n",
       "      <td>0.109</td>\n",
       "      <td>-0.001</td>\n",
       "      <td>-0.228</td>\n",
       "      <td>-0.073</td>\n",
       "      <td>-0.172</td>\n",
       "      <td>-0.020</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP:0000568</th>\n",
       "      <td>0.087</td>\n",
       "      <td>-0.016</td>\n",
       "      <td>-0.254</td>\n",
       "      <td>-0.052</td>\n",
       "      <td>-0.169</td>\n",
       "      <td>0.053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP:0000478</th>\n",
       "      <td>0.105</td>\n",
       "      <td>-0.047</td>\n",
       "      <td>-0.241</td>\n",
       "      <td>-0.033</td>\n",
       "      <td>-0.142</td>\n",
       "      <td>0.003</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>HP:0001250</th>\n",
       "      <td>0.064</td>\n",
       "      <td>-0.076</td>\n",
       "      <td>-0.236</td>\n",
       "      <td>-0.020</td>\n",
       "      <td>-0.005</td>\n",
       "      <td>0.030</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                0      1      2      3      4      5\n",
       "id                                                  \n",
       "HP:0001159  0.096 -0.007 -0.201 -0.119 -0.137  0.003\n",
       "HP:0006101  0.064 -0.042 -0.216 -0.085 -0.159  0.000\n",
       "HP:0001770  0.091  0.001 -0.189 -0.122 -0.142 -0.009\n",
       "HP:0001161  0.068 -0.057 -0.183 -0.096 -0.152 -0.024\n",
       "HP:0001829  0.084 -0.004 -0.178 -0.112 -0.172 -0.007\n",
       "HP:0001155  0.108 -0.077 -0.241 -0.038 -0.166 -0.016\n",
       "HP:0001760  0.109 -0.001 -0.228 -0.073 -0.172 -0.020\n",
       "HP:0000568  0.087 -0.016 -0.254 -0.052 -0.169  0.053\n",
       "HP:0000478  0.105 -0.047 -0.241 -0.033 -0.142  0.003\n",
       "HP:0001250  0.064 -0.076 -0.236 -0.020 -0.005  0.030"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ols.embeddings_dataframe(hand_terms, model=\"text-embedding-3-large_pca512\").iloc[:, :6].round(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8872d8b4",
   "metadata": {},
   "source": [
    "## Pairwise similarity\n",
    "\n",
    "Scores are cosine similarities computed locally from the vectors. OLS's own similarity\n",
    "endpoints report `(1 + cosine) / 2`; OAK converts these back to cosine so that all\n",
    "scores are on the same scale\n",
    "([#915](https://github.com/INCATools/ontology-access-kit/issues/915)).\n",
    "\n",
    "Different models disagree substantially, even on the same pair of terms:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "b07927ae",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:19.537121Z",
     "iopub.status.busy": "2026-10-05T01:10:19.536825Z",
     "iopub.status.idle": "2026-10-05T01:10:23.041144Z",
     "shell.execute_reply": "2026-10-05T01:10:23.039516Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>model</th>\n",
       "      <th>harrier-oss-v1-27b_pca512</th>\n",
       "      <th>llama-embed-nemotron-8b_pca512</th>\n",
       "      <th>text-embedding-3-large_pca512</th>\n",
       "      <th>text-embedding-3-small_pca512</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pair</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Finger syndactyly / Toe syndactyly</th>\n",
       "      <td>0.853</td>\n",
       "      <td>0.906</td>\n",
       "      <td>0.846</td>\n",
       "      <td>0.854</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Syndactyly / Finger syndactyly</th>\n",
       "      <td>0.939</td>\n",
       "      <td>0.875</td>\n",
       "      <td>0.922</td>\n",
       "      <td>0.871</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Syndactyly / Hand polydactyly</th>\n",
       "      <td>0.806</td>\n",
       "      <td>0.648</td>\n",
       "      <td>0.815</td>\n",
       "      <td>0.724</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Syndactyly / Seizure</th>\n",
       "      <td>0.369</td>\n",
       "      <td>0.366</td>\n",
       "      <td>0.268</td>\n",
       "      <td>0.392</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "model                               harrier-oss-v1-27b_pca512  \\\n",
       "pair                                                            \n",
       "Finger syndactyly / Toe syndactyly                      0.853   \n",
       "Syndactyly / Finger syndactyly                          0.939   \n",
       "Syndactyly / Hand polydactyly                           0.806   \n",
       "Syndactyly / Seizure                                    0.369   \n",
       "\n",
       "model                               llama-embed-nemotron-8b_pca512  \\\n",
       "pair                                                                 \n",
       "Finger syndactyly / Toe syndactyly                           0.906   \n",
       "Syndactyly / Finger syndactyly                               0.875   \n",
       "Syndactyly / Hand polydactyly                                0.648   \n",
       "Syndactyly / Seizure                                         0.366   \n",
       "\n",
       "model                               text-embedding-3-large_pca512  \\\n",
       "pair                                                                \n",
       "Finger syndactyly / Toe syndactyly                          0.846   \n",
       "Syndactyly / Finger syndactyly                              0.922   \n",
       "Syndactyly / Hand polydactyly                               0.815   \n",
       "Syndactyly / Seizure                                        0.268   \n",
       "\n",
       "model                               text-embedding-3-small_pca512  \n",
       "pair                                                               \n",
       "Finger syndactyly / Toe syndactyly                          0.854  \n",
       "Syndactyly / Finger syndactyly                              0.871  \n",
       "Syndactyly / Hand polydactyly                               0.724  \n",
       "Syndactyly / Seizure                                        0.392  "
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "\n",
    "pairs = [\n",
    "    (\"HP:0001159\", \"HP:0006101\"),  # syndactyly vs finger syndactyly (subclass)\n",
    "    (\"HP:0006101\", \"HP:0001770\"),  # finger vs toe syndactyly (siblings)\n",
    "    (\"HP:0001159\", \"HP:0001161\"),  # syndactyly vs hand polydactyly\n",
    "    (\"HP:0001159\", \"HP:0001250\"),  # syndactyly vs seizure (unrelated)\n",
    "]\n",
    "labels = dict(ols.labels({t for p in pairs for t in p}))\n",
    "rows = []\n",
    "for model in ols.embedding_models():\n",
    "    for s, o in pairs:\n",
    "        rows.append({\"model\": model, \"pair\": f\"{labels[s]} / {labels[o]}\",\n",
    "                     \"cosine\": ols.embedding_similarity(s, o, model=model)})\n",
    "pd.DataFrame(rows).pivot(index=\"pair\", columns=\"model\", values=\"cosine\").round(3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11ffdbad",
   "metadata": {},
   "source": [
    "Note that absolute cosine values are not comparable across models: each model has its\n",
    "own \"background\" similarity for unrelated terms. Rankings within a model are what matter.\n",
    "\n",
    "## Nearest neighbours and text search\n",
    "\n",
    "`nearest_entities` uses the OLS vector index; results can come from any ontology in OLS\n",
    "(useful for finding mappings). Obsolete classes are filtered out."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "b1f7926a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:23.044108Z",
     "iopub.status.busy": "2026-10-05T01:10:23.043823Z",
     "iopub.status.idle": "2026-10-05T01:10:26.007153Z",
     "shell.execute_reply": "2026-10-05T01:10:26.005610Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>label</th>\n",
       "      <th>cosine</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HP:0006101</td>\n",
       "      <td>Finger syndactyly</td>\n",
       "      <td>0.922</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HP:0001770</td>\n",
       "      <td>Toe syndactyly</td>\n",
       "      <td>0.909</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HP:0010554</td>\n",
       "      <td>Cutaneous finger syndactyly</td>\n",
       "      <td>0.868</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HP:0010704</td>\n",
       "      <td>1-2 finger cutaneous syndactyly</td>\n",
       "      <td>0.859</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HP:0010709</td>\n",
       "      <td>2-4 finger cutaneous syndactyly</td>\n",
       "      <td>0.857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>HP:0001233</td>\n",
       "      <td>2-3 finger cutaneous syndactyly</td>\n",
       "      <td>0.850</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>HP:0010713</td>\n",
       "      <td>1-5 toe syndactyly</td>\n",
       "      <td>0.844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>HP:0006097</td>\n",
       "      <td>3-4 finger osseus syndactyly</td>\n",
       "      <td>0.842</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           id                            label  cosine\n",
       "0  HP:0006101                Finger syndactyly   0.922\n",
       "1  HP:0001770                   Toe syndactyly   0.909\n",
       "2  HP:0010554      Cutaneous finger syndactyly   0.868\n",
       "3  HP:0010704  1-2 finger cutaneous syndactyly   0.859\n",
       "4  HP:0010709  2-4 finger cutaneous syndactyly   0.857\n",
       "5  HP:0001233  2-3 finger cutaneous syndactyly   0.850\n",
       "6  HP:0010713               1-5 toe syndactyly   0.844\n",
       "7  HP:0006097     3-4 finger osseus syndactyly   0.842"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model = \"text-embedding-3-large_pca512\"\n",
    "pd.DataFrame(\n",
    "    [(c, ols.label(c), round(score, 3)) for c, score in ols.nearest_entities(\"HP:0001159\", limit=8, model=model)],\n",
    "    columns=[\"id\", \"label\", \"cosine\"],\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "095e2054",
   "metadata": {},
   "source": [
    "Free-text queries are embedded by OLS. Only some OLS models can embed text; by default\n",
    "the first such model is used, and results are restricted to the adapter's ontology:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "7cdf3a70",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:26.010050Z",
     "iopub.status.busy": "2026-10-05T01:10:26.009698Z",
     "iopub.status.idle": "2026-10-05T01:10:41.533429Z",
     "shell.execute_reply": "2026-10-05T01:10:41.532181Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>label</th>\n",
       "      <th>cosine</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>HP:0001159</td>\n",
       "      <td>Syndactyly</td>\n",
       "      <td>0.821</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>HP:0010704</td>\n",
       "      <td>1-2 finger cutaneous syndactyly</td>\n",
       "      <td>0.813</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>HP:0006101</td>\n",
       "      <td>Finger syndactyly</td>\n",
       "      <td>0.711</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>HP:0010554</td>\n",
       "      <td>Cutaneous finger syndactyly</td>\n",
       "      <td>0.696</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>HP:0010709</td>\n",
       "      <td>2-4 finger cutaneous syndactyly</td>\n",
       "      <td>0.686</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           id                            label  cosine\n",
       "0  HP:0001159                       Syndactyly   0.821\n",
       "1  HP:0010704  1-2 finger cutaneous syndactyly   0.813\n",
       "2  HP:0006101                Finger syndactyly   0.711\n",
       "3  HP:0010554      Cutaneous finger syndactyly   0.696\n",
       "4  HP:0010709  2-4 finger cutaneous syndactyly   0.686"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(\n",
    "    [(c, ols.label(c), round(score, 3)) for c, score in ols.nearest_entities_to_text(\"webbed fingers\", limit=5)],\n",
    "    columns=[\"id\", \"label\", \"cosine\"],\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "90e2bd96",
   "metadata": {},
   "source": [
    "## Classic semantic similarity as vectors: the `closure` model\n",
    "\n",
    "Any adapter that can compute ancestors provides the `closure` model. Each term is a\n",
    "multi-hot vector over the union of the ancestors of the terms being encoded:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "0d1218b6",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:41.536503Z",
     "iopub.status.busy": "2026-10-05T01:10:41.536219Z",
     "iopub.status.idle": "2026-10-05T01:10:41.587141Z",
     "shell.execute_reply": "2026-10-05T01:10:41.585630Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(3, 16)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Abnormal foot morphology</th>\n",
       "      <th>Toe syndactyly</th>\n",
       "      <th>Abnormal toe morphology</th>\n",
       "      <th>Abnormality of the lower limb</th>\n",
       "      <th>Finger syndactyly</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Syndactyly</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Finger syndactyly</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Toe syndactyly</th>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                   Abnormal foot morphology  Toe syndactyly  \\\n",
       "Syndactyly                                0               0   \n",
       "Finger syndactyly                         0               0   \n",
       "Toe syndactyly                            1               1   \n",
       "\n",
       "                   Abnormal toe morphology  Abnormality of the lower limb  \\\n",
       "Syndactyly                               0                              0   \n",
       "Finger syndactyly                        0                              0   \n",
       "Toe syndactyly                           1                              1   \n",
       "\n",
       "                   Finger syndactyly  \n",
       "Syndactyly                         0  \n",
       "Finger syndactyly                  1  \n",
       "Toe syndactyly                     0  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from oaklib.utilities.embeddings.closure_embeddings import closure_embeddings\n",
    "\n",
    "hp = get_adapter(\"sqlite:obo:hp\")\n",
    "ids, vocab, m = closure_embeddings(hp, hand_terms[:3])\n",
    "closure_df = pd.DataFrame(m.astype(int), index=[hp.label(i) for i in ids], columns=[hp.label(v) for v in vocab])\n",
    "print(closure_df.shape)\n",
    "# show the dimensions (ancestors) that are not shared by all three terms\n",
    "closure_df.loc[:, closure_df.sum() < 3]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d820b885",
   "metadata": {},
   "source": [
    "Jaccard similarity over closure vectors is identical to the ancestor-set Jaccard computed\n",
    "by the `SemanticSimilarityInterface`:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "8d0976a3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:41.589825Z",
     "iopub.status.busy": "2026-10-05T01:10:41.589566Z",
     "iopub.status.idle": "2026-10-05T01:10:44.415589Z",
     "shell.execute_reply": "2026-10-05T01:10:44.413462Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          Syndactyly / Finger syndactyly    vector=0.9167 classic=0.9167\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   Finger syndactyly / Toe syndactyly       vector=0.6875 classic=0.6875\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          Syndactyly / Hand polydactyly     vector=0.4762 classic=0.4762\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "          Syndactyly / Seizure              vector=0.1429 classic=0.1429\n"
     ]
    }
   ],
   "source": [
    "from oaklib.datamodels.vocabulary import IS_A\n",
    "\n",
    "for s, o in pairs:\n",
    "    vec = hp.embedding_similarity(s, o, model=\"closure\", metric=\"jaccard\")\n",
    "    classic = hp.pairwise_similarity(s, o, predicates=[IS_A]).jaccard_similarity\n",
    "    print(f\"{labels[s]:>20} / {labels[o]:<20} vector={vec:.4f} classic={classic:.4f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "eb2090f8",
   "metadata": {},
   "source": [
    "## Comparing the similarity structure of models\n",
    "\n",
    "Below are all-by-all similarity matrices for the same ten terms: ontology closure\n",
    "(cosine) versus two OLS-hosted LLM embedding models. The closure model only sees the\n",
    "ontology graph; the LLM models only see the text of the terms."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "bd1c3002",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:44.418536Z",
     "iopub.status.busy": "2026-10-05T01:10:44.418228Z",
     "iopub.status.idle": "2026-10-05T01:10:48.656176Z",
     "shell.execute_reply": "2026-10-05T01:10:48.654270Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1760x572 with 4 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import numpy as np\n",
    "\n",
    "short = {t: ols.label(t) for t in hand_terms}\n",
    "panels = [\n",
    "    (\"closure (ontology)\", hp, \"closure\"),\n",
    "    (\"text-embedding-3-large\", ols, \"text-embedding-3-large_pca512\"),\n",
    "    (\"harrier-oss-v1-27b\", ols, \"harrier-oss-v1-27b_pca512\"),\n",
    "]\n",
    "fig, axes = plt.subplots(1, 3, figsize=(16, 5.2), sharey=True)\n",
    "for ax, (title, adapter, model) in zip(axes, panels):\n",
    "    m = adapter.embedding_similarity_matrix(hand_terms, model=model)\n",
    "    m = m.rename(index=short, columns=short)\n",
    "    sns.heatmap(m, ax=ax, cmap=SEQUENTIAL, vmin=0, vmax=1, square=True,\n",
    "                cbar=ax is axes[-1], cbar_kws={\"label\": \"cosine similarity\", \"shrink\": 0.8},\n",
    "                linewidths=1, linecolor=SURFACE)\n",
    "    ax.set_title(title)\n",
    "    ax.grid(False)\n",
    "    ax.tick_params(length=0)\n",
    "fig.tight_layout()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "183cf85d",
   "metadata": {},
   "source": [
    "All three put the limb terms in one block, and the eye terms and seizure apart from it.\n",
    "They differ in the details. The closure model scores the general terms (\"Abnormality of\n",
    "the hand\", \"Abnormal foot morphology\") as only moderately similar to their own descendants,\n",
    "because those descendants have many extra ancestors. The LLM models place them closer.\n",
    "The two LLM models also have very different \"background\" similarity between unrelated terms\n",
    "(compare the Seizure row). This is why absolute cosine thresholds do not transfer between\n",
    "models; rankings within a model are more meaningful.\n",
    "\n",
    "## Comparing sets of terms\n",
    "\n",
    "`embedding_termset_similarity` compares two sets of terms (e.g. a patient profile and a\n",
    "disease profile) by best-match average, returning the same `TermSetPairwiseSimilarity`\n",
    "object as classic semantic similarity:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "9b69ea2e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2026-10-05T01:10:48.659094Z",
     "iopub.status.busy": "2026-10-05T01:10:48.658756Z",
     "iopub.status.idle": "2026-10-05T01:10:48.669333Z",
     "shell.execute_reply": "2026-10-05T01:10:48.667900Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "best match average: 0.705\n",
      "  Finger syndactyly -> Syndactyly (0.922)\n",
      "  Microphthalmia -> Abnormality of the eye (0.703)\n"
     ]
    }
   ],
   "source": [
    "patient = [\"HP:0006101\", \"HP:0000568\"]           # finger syndactyly, microphthalmia\n",
    "disease = [\"HP:0001159\", \"HP:0000478\", \"HP:0001250\"]  # syndactyly, eye abnormality, seizure\n",
    "sim = ols.embedding_termset_similarity(patient, disease, model=\"text-embedding-3-large_pca512\", labels=True)\n",
    "print(\"best match average:\", round(sim.average_score, 3))\n",
    "for bm in sim.subject_best_matches.values():\n",
    "    print(f\"  {bm.match_source_label} -> {bm.match_target_label} ({bm.score:.3f})\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c5abd2ad",
   "metadata": {},
   "source": [
    "## LLM embeddings computed on demand\n",
    "\n",
    "The `llm:` adapter wraps another adapter and embeds term text with any model supported by\n",
    "the `llm` library (OpenAI, sentence-transformers via plugins, ...). The text template is\n",
    "configurable, so you can compare label-only with label + definition embeddings. This needs\n",
    "an API key (or a local model plugin), so it is not run here:\n",
    "\n",
    "```python\n",
    "llm_adapter = get_adapter(\"llm:sqlite:obo:hp\")\n",
    "llm_adapter.embedding_model_id = \"text-embedding-3-small\"\n",
    "llm_adapter.embedding_text_template = \"{label}: {definition}\"\n",
    "ids, matrix = llm_adapter.entity_embeddings(hand_terms)\n",
    "```\n",
    "\n",
    "## Command line\n",
    "\n",
    "The same operations are available from `runoak`:\n",
    "\n",
    "```bash\n",
    "runoak -i ols:hp embedding-models\n",
    "runoak -i ols:hp embeddings .desc//p=i HP:0001155 -m text-embedding-3-large_pca512 -o hand.tsv\n",
    "runoak -i ols:hp nearest-entities HP:0001159 -L 5\n",
    "runoak -i ols:hp nearest-entities --text \"webbed fingers\"\n",
    "runoak -i ols:hp embedding-similarity HP:0001159 HP:0006101 @ HP:0001770 HP:0001250\n",
    "runoak -i sqlite:obo:hp embedding-similarity -m closure --metric jaccard HP:0001159 @ HP:0001770\n",
    "```\n",
    "\n",
    "## See also\n",
    "\n",
    "- [Subsumption recapitulation](Subsumption-Recapitulation.ipynb): do LLM embeddings\n",
    "  encode the ontology hierarchy?\n",
    "- [Phenotype profile matching](Phenotype-Profile-Matching.ipynb): retrieving diseases from\n",
    "  noisy phenotype profiles with classic and embedding-based similarity."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}