@sjcrh/proteinpaint-server 2.117.0 → 2.118.0

package/src/serverconfig.js

@@ -61,7 +61,7 @@ if (!serverconfig.clustalo) serverconfig.clustalo = 'clustalo'
 if (!serverconfig.Rscript) serverconfig.Rscript = 'Rscript'
 if (!serverconfig.gfServer) serverconfig.gfServer = 'gfServer'
 if (!serverconfig.gfClient) serverconfig.gfClient = 'gfClient'
-if (!serverconfig.python) serverconfig.python = 'python3'
+// note: server/src/app.ts uses `setPythonBinPath()` from `@scjrch/proteinpaint-python`
 // NOTE: will set other cmd paths that require binpath after it's filled-in below
 
 /******************

package/utils/gsea.py

@@ -1,146 +0,0 @@
-# Test syntax: cat ~/sjpp/test.txt | time python gsea.py
-# test.txt contains the json string autogenerated by the commented out nodejs code.
-import blitzgsea as blitz  
-import json 
-import time  
-import sys  
-import sqlite3
-import random
-import os  
-import numpy as np  
-import pandas as pd  
-
-# Helper function to extract gene symbols from a dictionary
-def extract_symbols(x):
-    return x['symbol']  # Return the 'symbol' field from the dictionary
-
-# Main function
-try:
-    # Check if there is input from stdin
-    if sys.stdin.read(1):
-        # Read each line from stdin
-        for line in sys.stdin:
-            # Parse the JSON input
-            json_object = json.loads(line)
-            cachedir = json_object['cachedir']  # Get the cache directory from the JSON object
-            genes = json_object['genes']  # Get the genes from the JSON object
-            fold_change = json_object['fold_change']  # Get the fold change values from the JSON object
-            num_permutations = json_object['num_permutations'] # Number of permutations for GSEA analysis
-            table_name = json_object['geneset_group']  # Get the gene set group from the JSON object
-            filter_non_coding_genes = json_object['filter_non_coding_genes']  # Get the filter_non_coding_genes flag from the JSON object
-            db = json_object['db']  # Get the database path from the JSON object
-            # Create a DataFrame for the signature
-            df = {'Genes': genes, 'fold_change': fold_change}  # Create a dictionary with genes and fold change
-            signature = pd.DataFrame(df)  # Convert the dictionary to a DataFrame
-            
-            # Connect to the SQLite database
-            conn = sqlite3.connect(db)  # Connect to the SQLite database
-            cursor = conn.cursor()  # Create a cursor object
-
-            msigdb_library = {}  # Initialize an empty dictionary for the gene set library
-            if table_name == "REACTOME--blitzgsea": # Parse from blitzgsea reactome library
-              msigdb_library = blitz.enrichr.get_library("Reactome_2022")
-            elif table_name == "KEGG--blitzgsea": # Parse from blitzgsea KEGG library
-              msigdb_library = blitz.enrichr.get_library("KEGG_2021_Human")
-            elif table_name == "WikiPathways--blitzgsea": # Parse from blitzgsea WikiPathways library
-              msigdb_library = blitz.enrichr.get_library("WikiPathways_2019_Human")              
-            else: # Use geneset groups from msigdb    
-              # Query to get gene set IDs
-              query = f"SELECT id FROM terms WHERE parent_id='{table_name}'"  # SQL query to get gene set IDs
-              cursor.execute(query)  # Execute the query
-              
-              # Fetch all gene set IDs
-              rows = cursor.fetchall()  # Fetch all rows from the executed query
-              
-              start_loop_time = time.time()  # Record the start time of the loop
-              
-              # Iterate over gene set IDs and fetch corresponding genes
-              for row in rows:
-                  query2 = f"SELECT genes FROM term2genes WHERE id='{row[0]}'"  # SQL query to get genes for a gene set ID
-                  cursor.execute(query2)  # Execute the query
-                  rows2 = cursor.fetchall()  # Fetch all rows from the executed query
-                  row3 = json.loads(rows2[0][0])  # Parse the JSON data
-                  msigdb_library[row[0]] = list(set(map(extract_symbols, row3)))  # Extract only unique gene symbols and add them to the library. "set" command selects only unique genes 
-              #print ("msigdb_library:",msigdb_library) 
-                  
-              # Close the cursor and connection to the database
-              cursor.close()  # Close the cursor
-              conn.close()  # Close the connection
-              
-              stop_loop_time = time.time()  # Record the stop time of the loop
-              execution_time = stop_loop_time - start_loop_time  # Calculate the execution time
-              print(f"Execution time: {execution_time} seconds")  # Print the execution time
-
-            # Filter out non-coding genes if specified
-            if filter_non_coding_genes:
-                coding_genes_query = "SELECT * FROM codingGenes"  # SQL query to get coding genes
-                genedb = json_object['genedb']  # Get the gene database path from the JSON object
-                gene_conn = sqlite3.connect(genedb)  # Connect to the gene database
-                gene_cursor = gene_conn.cursor()  # Create a cursor object for the gene database
-                gene_cursor.execute(coding_genes_query)  # Execute the query to get coding genes
-                coding_genes_list = gene_cursor.fetchall()  # Fetch all coding genes
-                coding_genes_list = list(map(lambda x: x[0], coding_genes_list))  # Extract the gene symbols
-                signature = signature[signature['Genes'].isin(coding_genes_list)]  # Filter the signature to include only coding genes
-            
-            try:
-                # Check if geneset_name and pickle_file are present for generating the plot
-                geneset_name = json_object['geneset_name']  # Get the gene set name from the JSON object
-                pickle_file = json_object['pickle_file']  # Get the pickle file name from the JSON object
-                if os.path.isfile(os.path.join(cachedir, pickle_file)): # Check if the pickle file exists as it may not be in the same server that did the original GSEA computation
-                    result = pd.read_pickle(os.path.join(cachedir, pickle_file))  # Load the result from the pickle file
-                    fig = blitz.plot.running_sum(signature, geneset_name, msigdb_library, result=result.T, compact=True)  # Generate the running sum plot
-                else: # If pickle file is not found, redo the GSEA computation from scratch
-                    result = blitz.gsea(signature, msigdb_library, permutations=num_permutations).T  # Perform GSEA computation and transpose the result
-                    fig = blitz.plot.running_sum(signature, geneset_name, msigdb_library, result=result.T, compact=True)  # Generate the running sum plot
-                    result.to_pickle(os.path.join(cachedir, pickle_file))  # Save the result to a pickle file with same name
-                random_num = np.random.rand()  # Generate a random number for unique png filename
-                png_filename = f"gsea_plot_{random_num}.png"  # Create a filename for the plot
-                fig.savefig(os.path.join(cachedir, png_filename), bbox_inches='tight')  # Save the plot as a PNG file
-                print(f'image: {{"image_file": "{png_filename}"}}')  # Print the image file path in JSON format
-            except KeyError:
-                # Initial GSEA calculation and save the result to a pickle file
-                start_gsea_time = time.time()  # Record the start time of GSEA
-                if __name__ == "__main__":
-                    result = blitz.gsea(signature, msigdb_library, permutations=num_permutations).T  # Perform GSEA computation and transpose the result
-                    random_num = random.random() + time.time()  # Generate a random number for unique pickle filename
-                    pickle_filename = f"gsea_result_{random_num}.pkl"  # Create a filename for the pickle file
-                    result.to_pickle(os.path.join(cachedir, pickle_filename))  # Save the result to the pickle file
-                    gsea_str = f'{{"data": {result.to_json()}}}'  # Convert the result to JSON format
-                    pickle_str = f'{{"pickle_file": "{pickle_filename}"}}'  # Create a JSON string for the pickle file
-                    gsea_dict = json.loads(gsea_str)  # Parse the JSON string
-                    pickle_dict = json.loads(pickle_str)  # Parse the JSON string
-                    result_dict = {**gsea_dict, **pickle_dict}  # Merge the dictionaries
-                    print(f"result: {json.dumps(result_dict)}")  # Print the result in JSON format
-                stop_gsea_time = time.time()  # Record the stop time of GSEA
-                gsea_time = stop_gsea_time - start_gsea_time  # Calculate the GSEA execution time
-                print(f"GSEA time: {gsea_time} seconds")  # Print the GSEA execution time
-    else:
-        pass  # Do nothing if there is no input from stdin
-except (EOFError, IOError):
-    pass  # Handle EOFError and IOError exceptions gracefully
-
-# Function to extract plot data for GSEA visualization (NOT currently being used, but will be used for generating client side gsea plots)
-def extract_plot_data(signature, geneset, library, result, center=True):
-     print("signature", signature)
-     print("result", result)
-     print("geneset", geneset)
-     print("library", library)
-     signature = signature.copy()  # Create a copy of the signature DataFrame
-     signature.columns = ["i", "v"]  # Rename columns to 'i' and 'v'
-     signature = signature.sort_values("v", ascending=False).set_index("i")  # Sort by 'v' in descending order and set 'i' as index
-     signature = signature[~signature.index.duplicated(keep='first')]  # Remove duplicate indices, keeping the first occurrence
-   
-     if center:
-         signature.loc[:, "v"] -= np.mean(signature.loc[:, "v"])  # Center the signature values by subtracting the mean
-   
-     signature_map = {h: i for i, h in enumerate(signature.index)}  # Create a mapping of signature indices
-   
-     gs = set(library[geneset])  # Get the gene set from the library
-     hits = [i for i, x in enumerate(signature.index) if x in gs]  # Find the indices of hits in the signature
-   
-     running_sum, es = blitz.enrichment_score(np.array(np.abs(signature.iloc[:, 0])), signature_map, gs)  # Compute running sum and enrichment score
-     running_sum = list(running_sum)  # Convert running sum to a list
-     nn = np.where(np.abs(running_sum) == np.max(np.abs(running_sum)))[0][0]  # Find the index of the maximum absolute running sum
-   
-     running_sum_str = [str(elem) for elem in running_sum]  # Convert running sum elements to strings
-     print(f'result: {{"nn": {nn}, "running_sum": "{",".join(running_sum_str)}", "es": {es}}}')  # Print the result in JSON format

package/utils/plotBrainImaging.py

@@ -1,112 +0,0 @@
-import nibabel as nib  # Library for loading data from neuroimaging file formats such as NIfTI
-import matplotlib.pyplot as plt
-import matplotlib.colors as mcolors
-import matplotlib.cm as cm
-import numpy as np
-import sys
-import io
-import json
-
-if len(sys.argv) <= 1:
-	print('python3 '+sys.argv[0]+' <path/to/template/file> plane index filesJson.\n filesJson: dictionary containg sample files and color per category).\nplane: L (left, sagittal), F (front, coronal), T (top, axial)')
-	sys.exit(1)
-
-plane = sys.argv[2]
-if(plane != 'L' and plane != 'F' and plane != 'T'):
-	print('Invalid plane')
-	sys.exit(1)
-index = sys.argv[3]
-
-if(len(index) == 0):
-	print('Need to provide index')
-	sys.exit(1)
-
-templateFile = sys.argv[1]
-
-# load data from nifti files 
-template = nib.load(templateFile).get_fdata()
-
-vmaxSamples = sys.argv[4]
-
-if(len(vmaxSamples) == 0):
-	print('Need to provide max samples for normalization')
-	sys.exit(1)
-vmaxSamples = int(vmaxSamples)
-
-sampleFiles = json.loads(sys.argv[5])
-
-
-index = int(index)
-#  (left, sagittal), f (front, coronal), t (top, axial) 
-if plane == 'L':
-	slice = template[index,:,:]
-	slice = np.rot90(slice)
-elif plane == 'F':
-	slice = template[:,index,:]
-	slice = np.flip(np.rot90(slice),axis=1)
-
-else:# plane == 'T'
-	slice = template[:,:,index]
-	slice = np.flip(np.rot90(slice),axis=1)
-
-fig, ax = plt.subplots(1, 1)
-ax.imshow(slice, 'gray', filternorm=False, vmin=0, vmax=100)
-
-for key, value in sampleFiles.items():
-	if(len(value["samples"]) == 0) :
-		continue
-	# Load all sample files
-	sample_data = [nib.load(file_path).get_fdata() for file_path in value["samples"]]
-
-	# Initialize the result array with zeros
-	labels = np.zeros_like(sample_data[0])
-
-	# Sum all sample data
-	for data in sample_data:
-		labels += data
-
-	labels = np.ma.masked_where(labels == 0, labels) # Mask labels where they are 0
-
-
-	index = int(index)
-	#  (left, sagittal), f (front, coronal), t (top, axial) 
-	if plane == 'L':
-		label = labels[index,:,:]
-		label = np.rot90(label)
-	elif plane == 'F':
-		label = labels[:,index,:]
-		label = np.flip(np.rot90(label),axis=1)
-	else:# plane == 'T'
-		label = labels[:,:,index]
-		label = np.flip(np.rot90(label),axis=1)
-		
-
-
-	# create three subplots for sagittal, coronal and axial plane
-	vmin = 0
-	vmax = 100
-	alpha = 0.6
-
-	color = value['color']
-	print(color)
-	cmap = mcolors.LinearSegmentedColormap.from_list('my_cmap', ['white', color])
-	ax.imshow(label, cmap, alpha=alpha, filternorm=False,vmin=0,vmax=vmaxSamples)
-	ax.axis('off')
-
-# Create the color bar
-# if showLegend == 1:
-# 	# Create a color bar without changing figure size
-# 	norm = mcolors.Normalize(vmin=0, vmax=vmaxSamples)
-# 	sm = plt.cm.ScalarMappable(cmap=cmap, norm=norm)
-
-# 	cbar = plt.colorbar(sm, ax=ax, orientation='vertical', fraction=0.01, pad=0.05, alpha=alpha)
-# 	cbar.set_label('Combined Intensity', color='white', fontsize=6, labelpad=-10)
-# 	cbar.ax.text(0.5, 1.0001, vmaxSamples, ha='center', va='bottom', transform=cbar.ax.transAxes, color='white', fontsize=6)
-# 	cbar.ax.text(0.5, -0.0001, 0, ha='center', va='top', transform=cbar.ax.transAxes, color='white', fontsize=6)
-
-# Output the image data to stdout
-buf = io.BytesIO()
-plt.savefig(buf, format='png', bbox_inches='tight', facecolor='k')
-buf.seek(0)
-sys.stdout.buffer.write(buf.getvalue())
-plt.close()