add paircounts

Author: cactuskid

foldtree2/notebooks/benchmarks/alphabet_Information_content_benchmark.ipynb

@@ -1487,7 +1487,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 22,
+   "execution_count": null,
    "id": "7d89d6d2",
    "metadata": {},
    "outputs": [],
@@ -1739,7 +1739,72 @@
     "\t\n",
     "\tencoder.train()\n",
     "\tdecoder.train()\n",
-    "\t\n",
+    "\n",
+    "\n",
+    "\tfig2 = plt.figure(figsize=(6, 4))\n",
+    "\t#ramachandran plot\n",
+    "\tif 'angles' in sample_out and sample_out['angles'] is not None:\n",
+    "\t\ttrue_angles = data_sample['bondangles'].x.detach().cpu().numpy()\n",
+    "\t\tpred_angles = sample_out['angles'].detach().cpu().numpy()\n",
+    "\t\taxs2 = fig2.add_subplot(1, 1, 1)\n",
+    "\t\taxs2.scatter(true_angles[:, 0], true_angles[:, 1], label='True', color='blue', alpha=0.5, s=10)\n",
+    "\t\taxs2.scatter(pred_angles[:, 0], pred_angles[:, 1], label='Pred', color='red', alpha=0.5, s=10)\n",
+    "\t\taxs2.set_xlabel('Phi (N-Ca)')\n",
+    "\t\taxs2.set_ylabel('Psi (Ca-C)')\n",
+    "\t\taxs2.set_title('Ramachandran Plot')\n",
+    "\t\taxs2.legend()\n",
+    "\t\tfig2.tight_layout()\n",
+    "\t\t\n",
+    "\n",
+    "\tfig3, axs3 = plt.subplots(1, 3, figsize=(6, 4))\n",
+    "\t#separate angles plot\n",
+    "\tif 'angles' in sample_out and sample_out['angles'] is not None:\n",
+    "\t\ttrue_angles = data_sample['bondangles'].x.detach().cpu().numpy()\n",
+    "\t\tpred_angles = sample_out['angles'].detach().cpu().numpy()\n",
+    "\t\tangle_names = ['N-Ca-C', 'Ca-C-N', 'C-N-Ca']\n",
+    "\t\tangle_colors = ['r', 'g', 'b']\n",
+    "\t\t\n",
+    "\t\tfor i in range(3):\n",
+    "\t\t\taxs3[i].plot(true_angles[:, i], label='True ' + angle_names[i], \n",
+    "\t\t\t\t\t\t   color=angle_colors[i], alpha=0.5)\n",
+    "\t\t\taxs3[i].plot(pred_angles[:, i], label='Pred ' + angle_names[i], \n",
+    "\t\t\t\t\t\t   color=angle_colors[i], linestyle='--', alpha=0.5)\n",
+    "\t\t\taxs3[i].set_title(angle_names[i])\n",
+    "\t\t\taxs3[i].set_xlabel('Residue Index')\n",
+    "\t\t\taxs3[i].set_ylabel('Angle (radians)')\n",
+    "\t\t\taxs3[i].legend()\n",
+    "\t\tfig3.tight_layout()\n",
+    "\t\n",
+    "\t#confusion matrix for ss\n",
+    "\tfig4 = plt.figure(figsize=(6, 4))\n",
+    "\tif 'ss' in sample_out and sample_out['ss_pred'] is not None:\n",
+    "\t\tfrom sklearn.metrics import confusion_matrix, ConfusionMatrixDisplay\n",
+    "\t\ttrue_ss = data_sample['ss'].x.argmax(dim=1).cpu().numpy()\n",
+    "\t\tpred_ss = sample_out['ss_pred'].argmax(dim=1).cpu().numpy()\n",
+    "\t\tcm = confusion_matrix(true_ss, pred_ss)\n",
+    "\t\tdisp = ConfusionMatrixDisplay(confusion_matrix=cm)\n",
+    "\t\tdisp.plot(ax=axs3[0], cmap='Blues')\n",
+    "\t\tfig4.set_title('SS Confusion Matrix')\n",
+    "\t\tfig4.set_xlabel('Predicted SS')\n",
+    "\t\tfig4.set_ylabel('True SS')\n",
+    "\t\tfig4.tight_layout()\n",
+    "\t\n",
+    "\t#confusion matrix for aa\n",
+    "\tfig5 = plt.figure(figsize=(6, 4))\n",
+    "\tif 'aa' in sample_out and sample_out['aa'] is not None:\n",
+    "\t\ttrue_aa = data_sample['AA'].x.argmax(dim=1).cpu().numpy()\n",
+    "\t\tpred_aa = sample_out['aa'].argmax(dim=1).cpu().numpy()\n",
+    "\t\tcm_aa = confusion_matrix(true_aa, pred_aa)\n",
+    "\t\tdisp_aa = ConfusionMatrixDisplay(confusion_matrix=cm_aa)\n",
+    "\t\tdisp_aa.plot(ax=axs3[1], cmap='Blues')\n",
+    "\t\tfig5.set_title('AA Confusion Matrix')\n",
+    "\t\tfig5.set_xlabel('Predicted AA')\n",
+    "\t\tfig5.set_ylabel('True AA')\n",
+    "\t\tfig5.tight_layout()\n",
+    "\n",
+    "\tsup_figs = [fig2, fig3, fig4, fig5]\n",
+    "\tmetrics_dict['sup_figs'] = sup_figs\n",
+    "\n",
     "\treturn fig, metrics_dict , sample_out"
    ]
   },

foldtree2/notebooks/experiments/test_monodecoders.ipynb

@@ -49,7 +49,8 @@ dependencies = [
   "statsmodels",
   "pydssp",
   "ete3",
-  "pyyaml"
+  "pyyaml",
+  "ProDy"
 ]
 
 [project.scripts]