Minor bug fixes

app/fn.py

@@ -12,14 +12,13 @@ import tempfile
 
 import pandas as pd
 from bokeh.models import NumberFormatter, BooleanFormatter, HTMLTemplateFormatter
-from bokeh.resources import INLINE
 import gradio as gr
 import pytz
 import panel as pn
 import seaborn as sns
 from markdown import markdown
 from rdkit import Chem, RDConfig
-from rdkit.Chem import Crippen, Descriptors, rdMolDescriptors, Lipinski, rdmolops, Draw
+from rdkit.Chem import Crippen, Descriptors, rdMolDescriptors, Lipinski, rdmolops
 import requests
 
 from app import static
@@ -425,29 +424,11 @@ def create_result_table_html(summary_df, result_info, opts=(), progress=gr.Progr
     image_zoom_formatter = HTMLTemplateFormatter(
         template='<img src="data:image/svg+xml,<%= value %>" alt="Molecule" class="zoom-img">'
     )
-    uniprot_id_formatter = HTMLTemplateFormatter(
-        template='<% if (value == value) { '  # Check if value is not NaN
-                 'if (/^[OPQ][0-9][A-Z0-9]{3}[0-9]|[A-NR-Z][0-9]([A-Z][A-Z0-9]{2}[0-9]){1,2}$/.test(value)) '
-        # Check if value is a valid UniProt ID
-                 '{ %><a href="https://www.uniprot.org/uniprotkb/<%= value %>" target="_blank"><%= value %></a><% '
-        # Else treat it as a sequence or other plain-text string, line-warping every 60 characters
-                 '} else { %><div style="white-space: pre-wrap;"><%= value.match(/.{1,60}/g).join("<br>") '
-                 '%></div><% } %><% } else { %><% } %>'  # Output empty string if value is NaN
-    )
-    pubchem_id_formatter = HTMLTemplateFormatter(
-        template='<% if (value == value) { '  # Check if value is not NaN
-                 '%><a href="https://pubchem.ncbi.nlm.nih.gov/#query=<%= value %>" '
-                 'target="_blank"><%= value %></a>'
-                 '<% } else { %><% } %>'  # Output empty string if value is NaN
-    )
     bool_formatters = {col: BooleanFormatter() for col in bool_cols}
     float_formatters = {col: NumberFormatter(format='0.000') for col in html_df.select_dtypes('floating').columns}
     other_formatters = {
         'Compound': image_zoom_formatter,
         # 'Scaffold': image_zoom_formatter,
-        # 'Target FASTA': {'type': 'textarea', 'width': 60},
-        'Target ID': uniprot_id_formatter,
-        # 'Compound ID': pubchem_id_formatter,  ## TODO: add link to click for adding mol to the viewer
         'Pose': {'type': 'molDisplayButtonFormatter'},
     }
     formatters = {**bool_formatters, **float_formatters, **other_formatters}

app/main.py

@@ -1,7 +1,6 @@
-import io
-
 import spaces
 
+import io
 import os
 import tempfile
 import uuid
@@ -11,6 +10,7 @@ from pathlib import Path
 from time import sleep, time
 import urllib.parse
 
+from dotenv import load_dotenv
 import torch
 from email_validator import validate_email, EmailNotValidError
 from Bio import SeqIO
@@ -19,15 +19,17 @@ from gradio_rangeslider import RangeSlider
 from omegaconf import OmegaConf
 import pandas as pd
 from rdkit import Chem
-from rdkit.Chem import PandasTools, Draw
+from rdkit.Chem import PandasTools, Draw, rdDepictor
 
 from inference import (read_fragment_library, process_fragment_library, extract_pockets,
                        dock_fragments, generate_linkers, select_fragment_pairs)
 from app import static, fn, db
 
 
+load_dotenv()
 Path(tempfile.gettempdir(), 'gradio').mkdir(exist_ok=True)
-gr.set_static_paths(paths=["data/", "results/", "app/"])
+SERVER_DATA_DIR = os.getenv('DATA', 'results')
+gr.set_static_paths(paths=["data/", SERVER_DATA_DIR, "app/"])
 job_db = db.init_job_db()
 os.chmod('./fpocket', 0o755)
 
@@ -82,6 +84,8 @@ It might take a few minutes up to a few hours depending on the input size and th
 You may keep the page open and wait for job completion, or close the page and revisit later to look up the job status 
 using the job id. You will also receive an email notification once the job is done.
 ''',
+                        pred_lookup_btn: gr.Button(visible=False),
+                        pred_lookup_stop_btn: gr.Button(visible=True),
                     }
                 if job['status'] == "COMPLETED":
                     stop = True
@@ -94,7 +98,9 @@ using the job id. You will also receive an email notification once the job is do
                     yield {
                         pred_lookup_status: msg,
                         tabs: gr.Tabs(selected='result'),
-                        result_state: job
+                        result_state: job,
+                        pred_lookup_btn: gr.Button(visible=True),
+                        pred_lookup_stop_btn: gr.Button(visible=False),
                     }
                 if job['status'] == "FAILED":
                     stop = True
@@ -118,7 +124,7 @@ using the job id. You will also receive an email notification once the job is do
                 yield {
                     pred_lookup_status: msg,
                     pred_lookup_btn: gr.Button(visible=stop),
-                    pred_lookup_stop_btn: gr.Button(visible=stop),
+                    pred_lookup_stop_btn: gr.Button(visible= not stop),
                 }
             sleep(5)
 
@@ -197,23 +203,22 @@ def dock_link(
     linker_frag_dist, linker_strategy, linker_n_mols, linker_size, linker_steps,
     job_info
 ):
-    update_info = {}
     job_id = job_info['id']
+    job_db.job_update(
+        job_id=job_id,
+        update_info={'status': 'RUNNING'},
+    )
     pocket_extract_method = job_info['pocket_extraction_method']
     pocket_path_dict = job_info['protein_pocket_files']
+    update_info = {}
 
     config = OmegaConf.load('configs/gen_fbdd_v1.yaml')
     device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
     print(f'Using device: {device}')
     date_time = datetime.now().strftime("%Y-%m-%d_%H-%M-%S")
-    out_dir = f'results/{date_time}_{job_id}'
+    out_dir = Path(SERVER_DATA_DIR, f'{date_time}_{job_id}')
     frag_lib['X2'] = prot
-    frag_lib['ID2'] = Path(prot).stem
-
-    job_db.job_update(
-        job_id=job_id,
-        update_info={'status': 'RUNNING'},
-    )
+    frag_lib['ID2'] = str(Path(prot).stem)
 
     try:
         docking_df = dock_fragments(
@@ -272,7 +277,7 @@ def dock_link(
         update_info = {
             'status': "COMPLETED",
             'error': None,
-            'output_dir': out_dir,
+            'output_dir': str(out_dir),
             'type': job_type,
         }
         # return {result_state: job_info | update_info, run_state: {}}
@@ -544,7 +549,7 @@ with gr.Blocks(theme=THEME, title='GenFBDD', css=static.CSS, delete_cache=(3600,
                             info="Your job ID is a UUID4 string that you receive after submitting a job on the "
                                  "page or in the email notification.")
                         pred_lookup_example = gr.Button('Example', elem_classes=['example'], scale=1)
-                        pred_lookup_btn = gr.Button(value='Lookup the Job Status', variant='primary', visible=True)
+                        pred_lookup_btn = gr.Button(value='Query Status', variant='primary', visible=True)
                         pred_lookup_stop_btn = gr.Button(value='Stop Tracking', variant='stop', visible=False)
                         pred_lookup_status = gr.Markdown("**Job Status**", container=True)
 
@@ -645,7 +650,7 @@ with gr.Blocks(theme=THEME, title='GenFBDD', css=static.CSS, delete_cache=(3600,
                 raise ValueError(f"Unsupported method: {method}")
             return {input_prot_file: gr.File(str(file), visible=True)}
         except Exception as e:
-            raise gr.Error(f"Query error: {str(e)}")
+            gr.Warning(f"Query error: {str(e)}")
 
     prot_query_btn.click(
         fn=pdb_query,
@@ -759,7 +764,7 @@ with gr.Blocks(theme=THEME, title='GenFBDD', css=static.CSS, delete_cache=(3600,
         outputs=loader_html,
     )
 
-    job_valid.success(
+    job_valid_success = job_valid.success(
         fn=lambda job: [job['id'], gr.Tabs(selected='job')],
         inputs=[run_state],
         outputs=[pred_lookup_id, tabs],
@@ -767,7 +772,8 @@ with gr.Blocks(theme=THEME, title='GenFBDD', css=static.CSS, delete_cache=(3600,
         lambda: '<div class="loader"></ div>',
         outputs=loader_html,
     )
-    auto_job_lookup = job_valid.success(
+
+    auto_job_lookup = job_valid_success.success(
         fn=query_job_status,
         inputs=pred_lookup_id,
         outputs=[pred_lookup_status, tabs, result_state, pred_lookup_btn, pred_lookup_stop_btn],
@@ -815,11 +821,13 @@ with gr.Blocks(theme=THEME, title='GenFBDD', css=static.CSS, delete_cache=(3600,
         else:
             raise gr.Error('Invalid result type')
 
+        rdDepictor.SetPreferCoordGen(True)
         draw_opts = Draw.rdMolDraw2D.MolDrawOptions()
         draw_opts.clearBackground = False
         draw_opts.bondLineWidth = 0.5
         draw_opts.explicitMethyl = True
         draw_opts.singleColourWedgeBonds = True
+        draw_opts.addStereoAnnotation = False
         draw_opts.useCDKAtomPalette()
         PandasTools.drawOptions = draw_opts
         PandasTools.molSize = (90, 56)
@@ -918,5 +926,8 @@ with gr.Blocks(theme=THEME, title='GenFBDD', css=static.CSS, delete_cache=(3600,
 demo.launch(
     server_name='0.0.0.0',
     max_file_size="5mb",
-    ssr_mode=False
+    ssr_mode=False,
+    share=True,
+    share_server_address="ciddr-lab.ac.cn:7000",
+    share_server_protocol="https"
 )

inference.py

@@ -443,15 +443,13 @@ def dock_fragments(
             # TODO Use index instead of confidence in filename
             if save_docking:
                 sample_df['ligand_conf_path'] = [
-                    Path(
-                        docking_out_dir, f"{df['name'].iloc[idx]}_{i}-confidence{confidence[i]:.2f}.sdf"
-                    ) for i in range(n_samples)
+                    f"{df['name'].iloc[idx]}_{i}-confidence{confidence[i]:.2f}.sdf" for i in range(n_samples)
                 ]
             sample_df['ligand_mol']= [
                 create_mol_with_coords(
                     mol=RemoveAllHs(copy.deepcopy(lig)),
                     new_coords=pos,
-                    path=sample_df['ligand_conf_path'].iloc[i] if save_docking else None
+                    path=Path(docking_out_dir, sample_df['ligand_conf_path'].iloc[i]) if save_docking else None
                 ) for i, pos in enumerate(ligand_pos)
             ]
             # sample_df['ligand_pos'] = list(ligand_pos)

requirements.txt

@@ -17,6 +17,7 @@ hydra-core
 pandas
 pandarallel
 panel
+python-dotenv
 seaborn
 apscheduler
 tinydb

results/job_db.json

@@ -1,15 +1 @@
-{
-  "_default": {
-    "1": {
-      "id": "80cf2658-7a1c-48d6-8372-61b978177fe6",
-      "type": "linking",
-      "status": "COMPLETED",
-      "output_dir": "results/2025-01-01_01-01-01_80cf2658-7a1c-48d6-8372-61b978177fe6",
-      "protein_structure_file": "results/2025-01-01_01-01-01_80cf2658-7a1c-48d6-8372-61b978177fe6/1xkk.pdb",
-      "start_time": 1735693261,
-      "end_time": 1735693321,
-      "expiry_time": 1893459721,
-      "ip": "1.1.1.1"
-    }
-  }
-}
+{"_default": {"1": {"id": "80cf2658-7a1c-48d6-8372-61b978177fe6", "type": "linking", "status": "COMPLETED", "output_dir": "results/2025-01-01_01-01-01_80cf2658-7a1c-48d6-8372-61b978177fe6", "protein_structure_file": "results/2025-01-01_01-01-01_80cf2658-7a1c-48d6-8372-61b978177fe6/1xkk.pdb", "start_time": 1735693261, "end_time": 1735693321, "expiry_time": 1893459721, "ip": "1.1.1.1"}}}