Update app.py

app.py

@@ -2,6 +2,7 @@ import os
 import urllib.request
 import tarfile
 import shutil
+import pandas as pd
 from flask import Flask, render_template, request, jsonify, url_for
 from werkzeug.utils import secure_filename
 
@@ -20,7 +21,8 @@ import scAnalysis.differential as diff
 import scAnalysis.enrichment as enrich
 import scAnalysis.visualization as vis
 import scAnalysis.interactive_viz as iviz
-import scAnalysis.imputation as imp  # <-- ADDED IMPUTATION MODULE
+import scAnalysis.imputation as imp
+import scAnalysis.grn_inference as grn
 
 app = Flask(__name__)
 app.config['UPLOAD_FOLDER'] = './static/uploads'
@@ -106,12 +108,19 @@ def run_pipeline():
         elif norm_method == 'sctransform':
             pp.normalize_sctransform(data)
 
-        # <-- ADDED IMPUTATION EXECUTION BLOCK -->
+        # <-- EXPANDED IMPUTATION EXECUTION BLOCK -->
         if request.form.get('run_imputation') == 'true':
-            imp_k = int(request.form.get('imp_k', 7))
-            imp_thresh = float(request.form.get('imp_thresh', 0.72))
+            imp_method = request.form.get('imp_method', 'wnid')
             imp_pcs = int(request.form.get('imp_pcs', 30))
-            imp.impute_wnid(data, k=imp_k, dropout_thresh=imp_thresh, n_pcs=imp_pcs)
+            imp_k = int(request.form.get('imp_k', 7))
+            
+            if imp_method == 'wnid':
+                imp_thresh = float(request.form.get('imp_thresh', 0.72))
+                imp.impute_wnid(data, k=imp_k, dropout_thresh=imp_thresh, n_pcs=imp_pcs)
+            elif imp_method == 'knn':
+                imp.impute_knn_smooth(data, k=imp_k, n_pcs=imp_pcs)
+            elif imp_method == 'diffusion':
+                imp.impute_diffusion(data, t=3, n_pcs=imp_pcs, use_prebuilt_graph=False)
 
         organism = request.form.get('organism', 'human')
         cc.score_cell_cycle(data, organism=organism)
@@ -134,6 +143,7 @@ def run_pipeline():
         dim.run_pca(data, n_components=n_pcs)
         dim.neighbors(data, n_neighbors=n_neighbors, n_pcs=min(40, n_pcs))
 
+        # <-- BATCH CORRECTION -->
         batch_key = request.form.get('batch_key', '')
         if request.form.get('run_batch') == 'true' and batch_key in data.obs.columns:
             b_algo = request.form.get('batch_algo', 'harmony')
@@ -142,11 +152,24 @@ def run_pipeline():
             elif b_algo == 'combat': 
                 bc.combat(data, batch_key=batch_key)
             elif b_algo == 'mnn':
-                # MNN expects a list of SingleCellDatasets split by batch
                 batches = data.obs[batch_key].unique()
-                dataset_list = [data[data.obs[batch_key] == b] for b in batches]
+                dataset_list = [data[data.obs[batch_key] == b].copy() for b in batches]
                 data = bc.mnn_correct(dataset_list, batch_key=batch_key)
         
+        # <-- GRN INFERENCE -->
+        if request.form.get('run_grn') == 'true':
+            tf_input = request.form.get('tf_list', '')
+            tf_list = [tf.strip() for tf in tf_input.split(',') if tf.strip()]
+            if tf_list:
+                try:
+                    df_grn = grn.infer_grn_ridge(data, tf_list=tf_list, top_n_edges=5000)
+                    grn_path = os.path.join(res_dir, 'grn_edges.csv')
+                    df_grn.to_csv(grn_path, index=False)
+                    outputs.append({'type': 'file', 'url': url_for('static', filename='results/grn_edges.csv'), 'title': 'Download GRN Edges', 'icon': 'fa-project-diagram'})
+                except Exception as e:
+                    print("GRN Inference failed:", e)
+
+        # <-- DIMENSIONALITY REDUCTION -->
         if request.form.get('run_umap') == 'true':
             dim.run_umap(data, min_dist=float(request.form.get('umap_min_dist', 0.5)))
         if request.form.get('run_tsne') == 'true':
@@ -155,6 +178,7 @@ def run_pipeline():
             try: dim.run_phate(data)
             except: pass
 
+        # <-- CLUSTERING & TRAJECTORY -->
         clust_algo = request.form.get('clustering', 'leiden')
         res_k = float(request.form.get('resolution', 1.0))
         
@@ -173,6 +197,7 @@ def run_pipeline():
         diff_method = request.form.get('diff_method', 't-test')
         diff.rank_genes_groups(data, groupby='cluster', method=diff_method, use_raw=True)
 
+        # <-- VISUALIZATION -->
         if 'X_umap' in data.obsm:
             umap_path = os.path.join(res_dir, 'umap_clusters.png')
             vis.plot_umap(data, color="cluster", title="UMAP (Clusters)", save=umap_path)