Merge branch 'pandas_output' into pandas_output

EBjerrum · Mar 19, 2024 · 42673ff · 42673ff
2 parents cc17037 + 2973f70
commit 42673ff
Show file tree

Hide file tree

Showing 7 changed files with 535 additions and 252 deletions.
diff --git a/README.md b/README.md
@@ -71,6 +71,7 @@ There are a collection of notebooks in the notebooks directory which demonstrate
 * [Integrated hyperparameter tuning of Scikit-Learn estimator and Scikit-Mol transformer](https://github.com/EBjerrum/scikit-mol/blob/main/notebooks/06_hyperparameter_tuning.ipynb)
 * [Using parallel execution to speed up descriptor and fingerprint calculations](https://github.com/EBjerrum/scikit-mol/blob/main/notebooks/07_parallel_transforms.ipynb)
 * [Testing different fingerprints as part of the hyperparameter optimization](https://github.com/EBjerrum/scikit-mol/blob/main/notebooks/09_Combinatorial_Method_Usage_with_FingerPrint_Transformers.ipynb)
+* [Using pandas output for easy feature importance analysis and combine pre-exisitng values with new computations](https://github.com/EBjerrum/scikit-mol/blob/main/notebooks/10_pipeline_pandas_output.ipynb)
 
 
   We also put a software note on ChemRxiv. [https://doi.org/10.26434/chemrxiv-2023-fzqwd](https://doi.org/10.26434/chemrxiv-2023-fzqwd) 
@@ -93,4 +94,4 @@ Probably still, please check issues at GitHub and report there
 * Adrien Chaton [@adrienchaton](https://github.com/adrienchaton)
 * [@VincentAlexanderScholz](https://github.com/VincentAlexanderScholz)
 * [@RiesBen](https://github.com/RiesBen)
-
+* [@enricogandini](https://github.com/enricogandini)
diff --git a/notebooks/06_hyperparameter_tuning.ipynb b/notebooks/06_hyperparameter_tuning.ipynb
@@ -16,10 +16,10 @@
    "id": "7df4793c",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:35.701525Z",
-     "iopub.status.busy": "2023-03-19T08:54:35.700822Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.014139Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.013259Z"
+     "iopub.execute_input": "2024-03-18T13:17:06.835155Z",
+     "iopub.status.busy": "2024-03-18T13:17:06.834550Z",
+     "iopub.status.idle": "2024-03-18T13:17:07.938779Z",
+     "shell.execute_reply": "2024-03-18T13:17:07.938102Z"
     }
    },
    "outputs": [],
@@ -53,10 +53,10 @@
    "id": "45a8ebf1",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.017824Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.017350Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.022818Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.022080Z"
+     "iopub.execute_input": "2024-03-18T13:17:07.941632Z",
+     "iopub.status.busy": "2024-03-18T13:17:07.941344Z",
+     "iopub.status.idle": "2024-03-18T13:17:07.944976Z",
+     "shell.execute_reply": "2024-03-18T13:17:07.944438Z"
     }
    },
    "outputs": [],
@@ -87,10 +87,10 @@
    "id": "08c233a7",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.025520Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.025310Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.077231Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.076469Z"
+     "iopub.execute_input": "2024-03-18T13:17:07.947279Z",
+     "iopub.status.busy": "2024-03-18T13:17:07.947092Z",
+     "iopub.status.idle": "2024-03-18T13:17:07.986572Z",
+     "shell.execute_reply": "2024-03-18T13:17:07.985938Z"
     }
    },
    "outputs": [
@@ -124,17 +124,17 @@
    "id": "5363d05a",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.080021Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.079756Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.084574Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.084021Z"
+     "iopub.execute_input": "2024-03-18T13:17:07.989052Z",
+     "iopub.status.busy": "2024-03-18T13:17:07.988851Z",
+     "iopub.status.idle": "2024-03-18T13:17:07.993705Z",
+     "shell.execute_reply": "2024-03-18T13:17:07.993138Z"
     },
     "lines_to_next_cell": 2
    },
    "outputs": [],
    "source": [
     "\n",
-    "mol_list_train, mol_list_test, y_train, y_test = train_test_split(data.ROMol, data.pXC50, random_state=0)"
+    "mol_list_train, mol_list_test, y_train, y_test = train_test_split(data.ROMol, data.pXC50, random_state=42)"
    ]
   },
   {
@@ -151,10 +151,10 @@
    "id": "885daf12",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.087530Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.087275Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.454935Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.454153Z"
+     "iopub.execute_input": "2024-03-18T13:17:07.996298Z",
+     "iopub.status.busy": "2024-03-18T13:17:07.995898Z",
+     "iopub.status.idle": "2024-03-18T13:17:08.361248Z",
+     "shell.execute_reply": "2024-03-18T13:17:08.360550Z"
     }
    },
    "outputs": [],
@@ -182,10 +182,10 @@
    "id": "8fd14250",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.458340Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.458092Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.461663Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.461052Z"
+     "iopub.execute_input": "2024-03-18T13:17:08.364585Z",
+     "iopub.status.busy": "2024-03-18T13:17:08.364029Z",
+     "iopub.status.idle": "2024-03-18T13:17:08.367535Z",
+     "shell.execute_reply": "2024-03-18T13:17:08.366933Z"
     }
    },
    "outputs": [],
@@ -212,17 +212,18 @@
    "id": "fa082078",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.464188Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.463905Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.467187Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.466548Z"
+     "iopub.execute_input": "2024-03-18T13:17:08.369937Z",
+     "iopub.status.busy": "2024-03-18T13:17:08.369731Z",
+     "iopub.status.idle": "2024-03-18T13:17:08.372598Z",
+     "shell.execute_reply": "2024-03-18T13:17:08.372103Z"
     },
     "title": "Now hyperparameter tuning"
    },
    "outputs": [],
    "source": [
     "from sklearn.model_selection import RandomizedSearchCV\n",
-    "from sklearn.utils.fixes import loguniform"
+    "#from sklearn.utils.fixes import loguniform\n",
+    "from scipy.stats import loguniform"
    ]
   },
   {
@@ -239,18 +240,18 @@
    "id": "046e24d3",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.470133Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.469879Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.476370Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.475763Z"
+     "iopub.execute_input": "2024-03-18T13:17:08.374931Z",
+     "iopub.status.busy": "2024-03-18T13:17:08.374724Z",
+     "iopub.status.idle": "2024-03-18T13:17:08.380663Z",
+     "shell.execute_reply": "2024-03-18T13:17:08.380185Z"
     },
     "title": "Which keys do we have?"
    },
    "outputs": [
     {
      "data": {
       "text/plain": [
-       "dict_keys(['memory', 'steps', 'verbose', 'morgantransformer', 'ridge', 'morgantransformer__nBits', 'morgantransformer__parallel', 'morgantransformer__radius', 'morgantransformer__useBondTypes', 'morgantransformer__useChirality', 'morgantransformer__useCounts', 'morgantransformer__useFeatures', 'ridge__alpha', 'ridge__copy_X', 'ridge__fit_intercept', 'ridge__max_iter', 'ridge__positive', 'ridge__random_state', 'ridge__solver', 'ridge__tol'])"
+       "dict_keys(['memory', 'steps', 'verbose', 'morganfingerprinttransformer', 'ridge', 'morganfingerprinttransformer__nBits', 'morganfingerprinttransformer__parallel', 'morganfingerprinttransformer__radius', 'morganfingerprinttransformer__useBondTypes', 'morganfingerprinttransformer__useChirality', 'morganfingerprinttransformer__useCounts', 'morganfingerprinttransformer__useFeatures', 'ridge__alpha', 'ridge__copy_X', 'ridge__fit_intercept', 'ridge__max_iter', 'ridge__positive', 'ridge__random_state', 'ridge__solver', 'ridge__tol'])"
       ]
      },
      "execution_count": 8,
@@ -277,21 +278,21 @@
    "id": "cf2c45d7",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.478959Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.478683Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.483305Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.482688Z"
+     "iopub.execute_input": "2024-03-18T13:17:08.383218Z",
+     "iopub.status.busy": "2024-03-18T13:17:08.382844Z",
+     "iopub.status.idle": "2024-03-18T13:17:08.387097Z",
+     "shell.execute_reply": "2024-03-18T13:17:08.386532Z"
     },
     "lines_to_next_cell": 1
    },
    "outputs": [],
    "source": [
     "\n",
     "param_dist = {'ridge__alpha': loguniform(1e-2, 1e3),\n",
-    "            \"morgantransformer__nBits\": [256,512,1024,2048,4096],\n",
-    "            'morgantransformer__radius':[1,2,3,4],\n",
-    "            'morgantransformer__useCounts': [True,False],\n",
-    "            'morgantransformer__useFeatures':[True,False]}"
+    "            \"morganfingerprinttransformer__nBits\": [256,512,1024,2048,4096],\n",
+    "            'morganfingerprinttransformer__radius':[1,2,3,4],\n",
+    "            'morganfingerprinttransformer__useCounts': [True,False],\n",
+    "            'morganfingerprinttransformer__useFeatures':[True,False]}"
    ]
   },
   {
@@ -308,10 +309,10 @@
    "id": "fbb2cacd",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.485906Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.485645Z",
-     "iopub.status.idle": "2023-03-19T08:54:37.489973Z",
-     "shell.execute_reply": "2023-03-19T08:54:37.489382Z"
+     "iopub.execute_input": "2024-03-18T13:17:08.389557Z",
+     "iopub.status.busy": "2024-03-18T13:17:08.389341Z",
+     "iopub.status.idle": "2024-03-18T13:17:08.393214Z",
+     "shell.execute_reply": "2024-03-18T13:17:08.392741Z"
     },
     "title": "From https://scikit-learn.org/stable/auto_examples/model_selection/plot_randomized_search.html#sphx-glr-auto-examples-model-selection-plot-randomized-search-py"
    },
@@ -347,18 +348,18 @@
    "id": "bc66efa3",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:37.492677Z",
-     "iopub.status.busy": "2023-03-19T08:54:37.492327Z",
-     "iopub.status.idle": "2023-03-19T08:54:44.019837Z",
-     "shell.execute_reply": "2023-03-19T08:54:44.018820Z"
+     "iopub.execute_input": "2024-03-18T13:17:08.395641Z",
+     "iopub.status.busy": "2024-03-18T13:17:08.395386Z",
+     "iopub.status.idle": "2024-03-18T13:17:12.986572Z",
+     "shell.execute_reply": "2024-03-18T13:17:12.986010Z"
     }
    },
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Runtime: 6.52 for 25 iterations)\n"
+      "Runtime: 4.59 for 25 iterations)\n"
      ]
     }
    ],
@@ -380,10 +381,10 @@
    "id": "b2b3d623",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:44.026632Z",
-     "iopub.status.busy": "2023-03-19T08:54:44.024799Z",
-     "iopub.status.idle": "2023-03-19T08:54:44.034480Z",
-     "shell.execute_reply": "2023-03-19T08:54:44.033518Z"
+     "iopub.execute_input": "2024-03-18T13:17:12.989282Z",
+     "iopub.status.busy": "2024-03-18T13:17:12.988845Z",
+     "iopub.status.idle": "2024-03-18T13:17:12.992724Z",
+     "shell.execute_reply": "2024-03-18T13:17:12.992202Z"
     },
     "lines_to_next_cell": 0
    },
@@ -393,16 +394,16 @@
      "output_type": "stream",
      "text": [
       "Model with rank: 1\n",
-      "Mean validation score: 0.492 (std: 0.048)\n",
-      "Parameters: {'morgantransformer__nBits': 2048, 'morgantransformer__radius': 4, 'morgantransformer__useCounts': False, 'morgantransformer__useFeatures': False, 'ridge__alpha': 18.49687467370123}\n",
+      "Mean validation score: 0.526 (std: 0.183)\n",
+      "Parameters: {'morganfingerprinttransformer__nBits': 4096, 'morganfingerprinttransformer__radius': 2, 'morganfingerprinttransformer__useCounts': False, 'morganfingerprinttransformer__useFeatures': False, 'ridge__alpha': 0.03506956723396942}\n",
       "\n",
       "Model with rank: 2\n",
-      "Mean validation score: 0.446 (std: 0.188)\n",
-      "Parameters: {'morgantransformer__nBits': 256, 'morgantransformer__radius': 2, 'morgantransformer__useCounts': False, 'morgantransformer__useFeatures': False, 'ridge__alpha': 0.22311472368521185}\n",
+      "Mean validation score: 0.523 (std: 0.127)\n",
+      "Parameters: {'morganfingerprinttransformer__nBits': 4096, 'morganfingerprinttransformer__radius': 4, 'morganfingerprinttransformer__useCounts': False, 'morganfingerprinttransformer__useFeatures': False, 'ridge__alpha': 7.715592595566691}\n",
       "\n",
       "Model with rank: 3\n",
-      "Mean validation score: 0.440 (std: 0.025)\n",
-      "Parameters: {'morgantransformer__nBits': 4096, 'morgantransformer__radius': 3, 'morgantransformer__useCounts': False, 'morgantransformer__useFeatures': True, 'ridge__alpha': 6.023414585108017}\n",
+      "Mean validation score: 0.495 (std: 0.096)\n",
+      "Parameters: {'morganfingerprinttransformer__nBits': 2048, 'morganfingerprinttransformer__radius': 3, 'morganfingerprinttransformer__useCounts': False, 'morganfingerprinttransformer__useFeatures': False, 'ridge__alpha': 28.390132035436164}\n",
       "\n"
      ]
     }
@@ -433,19 +434,19 @@
    "id": "cb369a0e",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:44.041562Z",
-     "iopub.status.busy": "2023-03-19T08:54:44.039729Z",
-     "iopub.status.idle": "2023-03-19T08:54:44.258048Z",
-     "shell.execute_reply": "2023-03-19T08:54:44.256978Z"
+     "iopub.execute_input": "2024-03-18T13:17:12.995596Z",
+     "iopub.status.busy": "2024-03-18T13:17:12.995274Z",
+     "iopub.status.idle": "2024-03-18T13:17:13.177121Z",
+     "shell.execute_reply": "2024-03-18T13:17:13.176195Z"
     }
    },
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "No Standardization 0.5222\n",
-      "With Standardization 0.5222\n"
+      "No Standardization 0.6791\n",
+      "With Standardization 0.6791\n"
      ]
     }
    ],
@@ -477,19 +478,25 @@
    "id": "f6426b23",
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-03-19T08:54:44.264895Z",
-     "iopub.status.busy": "2023-03-19T08:54:44.264460Z",
-     "iopub.status.idle": "2023-03-19T08:54:44.295917Z",
-     "shell.execute_reply": "2023-03-19T08:54:44.294841Z"
+     "iopub.execute_input": "2024-03-18T13:17:13.180039Z",
+     "iopub.status.busy": "2024-03-18T13:17:13.179823Z",
+     "iopub.status.idle": "2024-03-18T13:17:13.196436Z",
+     "shell.execute_reply": "2024-03-18T13:17:13.195835Z"
     }
    },
    "outputs": [
     {
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Predictions with no standardization: [6.29056823 6.31806374 6.34706809 6.44119607 6.41499794]\n",
-      "Predictions with standardization:    [6.29056823 6.29056823 6.29056823 6.29056823 6.29056823]\n"
+      "Predictions with no standardization: [5.77626555 5.94364787 5.94364787 6.13649679 6.04966392]\n"
+     ]
+    },
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Predictions with standardization:    [5.77626555 5.77626555 5.77626555 5.77626555 5.77626555]\n"
      ]
     }
    ],
@@ -539,7 +546,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.9"
+   "version": "3.10.12"
   }
  },
  "nbformat": 4,

diff --git a/notebooks/06_hyperparameter_tuning.py b/notebooks/06_hyperparameter_tuning.py
@@ -7,7 +7,7 @@
 #       extension: .py
 #       format_name: percent
 #       format_version: '1.3'
-#       jupytext_version: 1.14.5
+#       jupytext_version: 1.16.1
 #   kernelspec:
 #     display_name: Python 3.9.4 ('rdkit')
 #     language: python
@@ -96,7 +96,8 @@
 
 # %% Now hyperparameter tuning
 from sklearn.model_selection import RandomizedSearchCV
-from sklearn.utils.fixes import loguniform
+#from sklearn.utils.fixes import loguniform
+from scipy.stats import loguniform
 
 # %% [markdown]
 # With the pipelines, getting the names of the parameters to tune is a bit more tricky, as they are concatenations of the name of the step and the parameter with double underscores in between. We can get the available parameters from the pipeline with the get_params() method, and select the parameters we want to change from there.