using v6 data

sustainable-processes · Feb 5, 2024 · c3b89f9 · c3b89f9
1 parent 844eb12
commit c3b89f9
Showing 1 changed file with 33 additions and 11 deletions.
diff --git a/notebooks/dataset_freq.ipynb b/notebooks/dataset_freq.ipynb
@@ -134,53 +134,75 @@
      "output_type": "stream",
      "text": [
       "\n",
-      "        reactant & 317184 // 0 // 18.35 & \n",
+      "        reactant & 503625 // 0 // 12.96 & \n",
       "        \n",
       "\n",
-      "        product & 382850 // 0 // 0.0 & \n",
+      "        product & 694279 // 0 // 0.0 & \n",
       "        \n",
       "\n",
-      "        solvent & 85 // 313 // 28.02 & \n",
+      "        solvent & 104 // 316 // 45.72 & \n",
       "        \n",
       "\n",
-      "        agent & 255 // 11945 // 37.04 & \n",
+      "        agent & 275 // 24547 // 60.93 & \n",
       "        \n"
      ]
     }
    ],
    "source": [
-    "path = \"/Users/dsw46/Projects_local/orderly_07_06/ORDerly/data/orderly/uspto_no_trust/filtered/filtered_orderly_ord.parquet\"\n",
+    "# path = \"/Users/dsw46/Projects_local/orderly_07_06/ORDerly/data/orderly/uspto_no_trust/filtered/filtered_orderly_ord.parquet\"\n",
+    "path =   \"/Users/dsw46/Projects_local/ORDerly_jcim_response/ORDerly/data/orderly/uspto_no_trust/filtered/filtered_orderly_ord.parquet\"\n",
     "threshold = 100\n",
     "list_of_cols = [reactant_cols, product_cols, solv_cols, agent_cols]\n",
     "build_overleaf_table(path, list_of_cols, threshold)"
    ]
   },
   {
    "cell_type": "code",
-   "execution_count": 6,
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "['catalyst_000', 'reagent_000', 'reagent_001']"
+      ]
+     },
+     "execution_count": 7,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "catalyst_cols+reagent_cols"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
    "metadata": {},
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
       "\n",
-      "        reactant & 40020 // 0 // 25.7 & \n",
+      "        reactant & 207066 // 0 // 6.95 & \n",
       "        \n",
       "\n",
-      "        product & 38816 // 0 // 0.0 & \n",
+      "        product & 253908 // 0 // 0.0 & \n",
       "        \n",
       "\n",
-      "        solvent & 29 // 204 // 40.88 & \n",
+      "        solvent & 59 // 598 // 65.84 & \n",
       "        \n",
       "\n",
-      "        catalyst & 48 // 447 // 56.18 & \n",
+      "        catalyst & 50 // 546 // 92.54 & \n",
       "        \n"
      ]
     }
    ],
    "source": [
-    "path = \"/Users/dsw46/Projects_local/orderly_07_06/ORDerly/data/orderly/uspto_with_trust/filtered/filtered_orderly_ord.parquet\"\n",
+    "# path = \"/Users/dsw46/Projects_local/orderly_07_06/ORDerly/data/orderly/uspto_with_trust/filtered/filtered_orderly_ord.parquet\"\n",
+    "path = \"/Users/dsw46/Projects_local/ORDerly_jcim_response/ORDerly/data/orderly/uspto_with_trust/filtered/filtered_orderly_ord.parquet\"\n",
     "threshold = 100\n",
     "list_of_cols = [reactant_cols, product_cols, solv_cols, catalyst_cols+reagent_cols]\n",
     "build_overleaf_table(path, list_of_cols, threshold)"