last changes for CR version WWW2019

lisette-espin · Mar 14, 2019 · 80e3a89 · 80e3a89
1 parent 6ceee72
commit 80e3a89
Show file tree

Hide file tree

Showing 9 changed files with 3,231 additions and 5,475 deletions.
diff --git a/2. BioPortal (valid).ipynb b/2. BioPortal (valid).ipynb
diff --git a/3. BioPortal (khop-matrices-lcc).ipynb b/3. BioPortal (khop-matrices-lcc).ipynb
diff --git a/4. BioPortal (distance-matrices-lcc).ipynb b/4. BioPortal (distance-matrices-lcc).ipynb
@@ -22,7 +22,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 1,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -52,7 +52,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -104,7 +104,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 3,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -156,11 +156,65 @@
   },
   {
    "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 5,
    "metadata": {
-    "scrolled": false
+    "scrolled": true
    },
-   "outputs": [],
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "loading submissions...\n",
+      "2019-02-15 19:47:43\t/bigdata/lespin/datasets/bioportal/submissions.json loaded!\n",
+      "2019-02-15 19:47:43\t- 642 ontologies\n",
+      "2019-02-15 19:47:43\t- 13 years\n",
+      "loading ontologies...\n",
+      "generating distance matrices for 12 ontologies...\n",
+      "LCC_LOINC_2015_9HOP.npz\n",
+      "LCC_LOINC_2015_1HOP.npz\n",
+      "LCC_LOINC_2015_4HOP.npz\n",
+      "LCC_LOINC_2015_5HOP.npz\n",
+      "LCC_LOINC_2015_2HOP.npz\n",
+      "LCC_LOINC_2015_7HOP.npz\n",
+      "LCC_LOINC_2015_3HOP.npz\n",
+      "LCC_LOINC_2015_8HOP.npz\n",
+      "LCC_LOINC_2015_12HOP.npz\n",
+      "LCC_LOINC_2015_10HOP.npz\n",
+      "LCC_LOINC_2015_11HOP.npz\n",
+      "LCC_LOINC_2015_6HOP.npz\n",
+      "LCC_LOINC_2015_13HOP.npz\n",
+      "file:LCC_LOINC_2015_9HOP.npz, k:9\n",
+      "2019-02-15 19:47:44\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_9HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_1HOP.npz, k:1\n",
+      "2019-02-15 19:47:45\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_1HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_4HOP.npz, k:4\n",
+      "2019-02-15 19:49:46\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_4HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_5HOP.npz, k:5\n",
+      "2019-02-15 19:51:53\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_5HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_2HOP.npz, k:2\n",
+      "2019-02-15 19:53:42\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_2HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_7HOP.npz, k:7\n",
+      "2019-02-15 19:56:32\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_7HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_3HOP.npz, k:3\n",
+      "2019-02-15 19:59:38\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_3HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_8HOP.npz, k:8\n",
+      "2019-02-15 20:03:05\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_8HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_12HOP.npz, k:12\n",
+      "2019-02-15 20:06:05\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_12HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_10HOP.npz, k:10\n",
+      "2019-02-15 20:09:03\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_10HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_11HOP.npz, k:11\n",
+      "2019-02-15 20:11:42\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_11HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_6HOP.npz, k:6\n",
+      "2019-02-15 20:17:42\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_6HOP.npz loaded!\n",
+      "file:LCC_LOINC_2015_13HOP.npz, k:13\n",
+      "2019-02-15 20:26:11\t/bigdata/lespin/bioportal/ontologies/hops/LCC_LOINC_2015_13HOP.npz loaded!\n",
+      "2019-02-15 22:56:05\t/bigdata/lespin/bioportal/ontologies/distances/LCC_LOINC_2015_HOPs.npz saved!\n",
+      "max khop (max-diameter): 13\n"
+     ]
+    }
+   ],
    "source": [
     "print('loading submissions...')\n",
     "submissions_dict = get_submissions(SUBMISSIONS_FN)\n",
@@ -175,15 +229,23 @@
     "for index,row in tmp.sort_values(\"E'\", ascending=True).iterrows():\n",
     "    onto_name = row['ontology']\n",
     "    \n",
-    "    if onto_name in ['LOINC','SNOMEDCT']:\n",
+    "    if onto_name not in ['LOINC']:\n",
     "        continue\n",
     "\n",
     "    submission = submissions_dict[onto_name][YEAR]\n",
     "    onto = Ontology(name=onto_name, year=YEAR, submission_id=submission['submissionId'], root_folder=ONTOLOGY_ROOT)\n",
     "    m = onto.create_distance_matrix(path=ONTOLOGY_DISTANCE_OUTPUT, hopspath=ONTOLOGY_HOPS_OUTPUT, lcc=True)\n",
     "    print(\"max khop (max-diameter): {}\".format(m.max()))    \n",
-    "    del(m)\n",
-    "    break\n"
+    "    del(m)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "del(onto)"
    ]
   },
   {