README

README.md

@@ -3,7 +3,9 @@ The project aims at using state-of-the-art machine learning methods, and in part
 
 ## Data
 ### Raw
-- cs_45.json; cell line terms, extracted from Cellosaurus; 145673 terms
-- cs_pos_pmid_set.tsv; curated positive samples, extracted from Cellosaurus, 22719 PMIDs
-- gs_neg_pmid.tsv; curated negative samples, extracted from Google Scholar, 475 PMIDs
-- ls_neg_pmid.tsv; uncurated negative samples, extracted from LitSuggest, 645 PMIDs
+- cs_term_45.0; cell line terms, extracted from Cellosaurus, 145673 terms
+#### PMID
+- Cellosaurus; positive samples, extracted from Cellosaurus, 22719 PMIDs
+- CellosaurusAB; positive samples, extracted from Cellosaurus, curated, high portion of seminal papers, 10.000 PMIDs
+- GoogleScholar; negative samples, extracted from rejected Google Scholar results, curated, 509 PMIDs
+- LitSuggest; negative samples, extract from rejected LitSuggest results, 645 PMIDs

notebook/data_proc_pmid.ipynb

@@ -2,7 +2,7 @@
  "cells": [
   {
    "cell_type": "code",
-   "execution_count": 130,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -17,6 +17,7 @@
     "import random\n",
     "from munch import Munch\n",
     "import numpy as np\n",
+    "from tab\n",
     "\n",
     "with initialize(\n",
     "    version_base=None,\n",