feat: add a CLI entrypoint for upload_vcf

clintval · clintval · commit e2b2bc216418 · 2024-12-13T16:34:15.000-10:00
diff --git a/.gitignore b/.gitignore
@@ -1,4 +1,6 @@
+.DS_Store
 .vscode/
+testdata/
 
 # Byte-compiled / optimized / DLL files
 __pycache__/
diff --git a/README.md b/README.md
@@ -18,6 +18,57 @@ The package can be installed with `pip`:
 pip install tp53
 ```
 
+### Uploading a VCF to the Seshat Server
+
+Upload a VCF to the Seshat TP53 annotation server using a headless browser.
+
+```bash
+❯ python -m tp53.seshat.upload_vcf \
+      --input "input.vcf" \
+      --email "example@gmail.com"
+```
+
+This tool is used to programmatically configure and upload batch variants in VCF format to the Seshat annotation server.
+The tool works by building a headless Chrome browser instance and then interacting with the Seshat website directly through simulated key presses and mouse clicks.
+Unfortunately, Seshat does not provide a native programmatic API and one could not be reverse engineered.
+Seshat also utilizes custom JavaScript in their form processing, so a lightweight approach of simply interacting with the HTML form elements was also not possible.
+
+##### VCF Input Requirements
+
+Seshat will not let the user know why a VCF fails to annotate, but it has been observed that Seshat can fail to parse some of VarDictJava's structural variants (SVs) as valid variant records.
+One solution that has worked in the past is to remove SVs.
+The following command will exclude all variants with a non-empty SVTYPE INFO key:
+
+  bcftools view in.vcf --exclude 'SVTYPE!="."' > out.noSV.vcf
+
+##### Automation
+
+There are no terms and conditions posted on the Seshat annotation server's website, and there is no server-side `robots.txt` rule set.
+In lieu of usage terms, we strongly encourage all users of this script to respect the Seshat resource by adhering to the following best practice:
+
+  - Minimize Load: Limit the rate of requests to the server
+  - Minimize Connections: Limit the number of concurrent requests
+
+If you need to batch process dozens, or hundreds, of VCF callsets, you may consider improving this underlying Python script to randomize the user agent and IP address of your headless browser session to prevent from being labelled as a bot.
+
+##### Environment Setup
+
+This script relies on Google Chrome:
+
+```console
+brew install --cask google-chrome
+```
+
+Distributions of MacOS may require you to authenticate the Chrome driver:
+
+  - https://stackoverflow.com/a/60362134
+
+##### References
+
+> Soussi, Thierry, et al. “Recommendations for Analyzing and Reporting TP53
+> Gene Variants in the High-Throughput Sequencing Era.” Human Mutation,
+> vol. 35, no. 6, 2014, pp. 766–778., doi:10.1002/humu.22561.
+
 ## Development and Testing
 
 See the [contributing guide](./CONTRIBUTING.md) for more information.
diff --git a/poetry.lock b/poetry.lock
diff --git a/pyproject.toml b/pyproject.toml
@@ -31,6 +31,7 @@ classifiers   = [
 [tool.poetry.dependencies]
 python                   = "^3.11"
 beautifulsoup4           = "~4.12"
+chromedriver-py          = "*"
 google-api-python-client = "~2.151"
 google-auth-httplib2     = "~0.2"
 google-auth-oauthlib     = "~1.2.1"
@@ -123,7 +124,7 @@ exclude                     = [
 ]
 
 [[tool.mypy.overrides]]
-module = "defopt"
+module = "chromedriver_py"
 ignore_missing_imports = true
 
 [[tool.mypy.overrides]]
diff --git a/tests/seshat/test_upload.py b/tests/seshat/test_upload.py
@@ -1,4 +1,4 @@
-from tp53.seshat import HumanGenomeAssembly
+from tp53.seshat.upload_vcf import HumanGenomeAssembly
 
 
 def test_human_genome_assembly() -> None:
diff --git a/tp53/seshat/__init__.py b/tp53/seshat/__init__.py
@@ -1,4 +1 @@
 from tp53.seshat._exceptions import SeshatError as SeshatError
-from tp53.seshat._gmail_find import find_in_gmail as find_in_gmail
-from tp53.seshat._upload import HumanGenomeAssembly as HumanGenomeAssembly
-from tp53.seshat._upload import upload_vcf as upload_vcf
diff --git a/tp53/seshat/find_in_gmail/__init__.py b/tp53/seshat/find_in_gmail/__init__.py
@@ -0,0 +1 @@
+from ._find_in_gmail import find_in_gmail as find_in_gmail
diff --git a/tp53/seshat/find_in_gmail/__main__.py b/tp53/seshat/find_in_gmail/__main__.py
@@ -0,0 +1,2 @@
+if __name__ == "__main__":
+    ...
diff --git a/tp53/seshat/find_in_gmail/_find_in_gmail.py b/tp53/seshat/find_in_gmail/_find_in_gmail.py
@@ -20,7 +20,7 @@
 from google_auth_oauthlib.flow import InstalledAppFlow
 from googleapiclient.discovery import build as build_google_client
 
-from ._exceptions import SeshatError
+from .._exceptions import SeshatError
 
 logger: Logger = getLogger("tp53.seshat")
 
diff --git a/tp53/seshat/upload_vcf/__init__.py b/tp53/seshat/upload_vcf/__init__.py
@@ -0,0 +1,2 @@
+from ._upload_vcf import HumanGenomeAssembly as HumanGenomeAssembly
+from ._upload_vcf import upload_vcf as upload_vcf
diff --git a/tp53/seshat/upload_vcf/__main__.py b/tp53/seshat/upload_vcf/__main__.py
@@ -0,0 +1,117 @@
+"""
+Upload a VCF to the Seshat TP53 annotation server using a headless browser.
+
+This tool is used to programmatically configure and upload batch variants in VCF
+format to the Seshat annotation server. The tool works by building a headless
+Chrome browser instance and then interacting with the Seshat website directly
+through simulated key presses and mouse clicks. Unfortunately, Seshat does not
+provide a native programmatic API and one could not be reverse engineered.
+Seshat also utilizes custom JavaScript in their form processing, so a
+lightweight approach of simply interacting with the HTML form elements was
+also not possible.
+
+#### VCF Input Requirements
+
+Seshat will not let the user know why a VCF fails to annotate, but it has
+been observed that Seshat can fail to parse some of VarDictJava's structural
+variants (SVs) as valid variant records. One solution that has worked in the
+past is to remove SVs. The following command will exclude all variants with a
+non-empty SVTYPE INFO key:
+
+  bcftools view in.vcf --exclude 'SVTYPE!="."' > out.noSV.vcf
+
+#### Automation
+
+There are no terms and conditions posted on the Seshat annotation server's
+website, and there is no server-side `robots.txt` rule set. In lieu of usage
+terms, we strongly encourage all users of this script to respect the Seshat
+resource by adhering to the following best practice:
+
+  - Minimize Load: Limit the rate of requests to the server
+  - Minimize Connections: Limit the number of concurrent requests
+
+If you need to batch process dozens, or hundreds, of VCF callsets, you may
+consider improving this underlying Python script to randomize the user agent and
+IP address of your headless browser session to prevent from being labelled as a
+bot.
+
+#### Environment Setup
+
+This script relies on Chrome:
+
+  brew install --cask google-chrome
+
+Distributions of MacOS require you to authenticate the Chrome driver:
+
+  - https://stackoverflow.com/a/60362134
+
+#### References
+
+  1. Soussi, Thierry, et al. “Recommendations for Analyzing and Reporting TP53
+     Gene Variants in the High-Throughput Sequencing Era.” Human Mutation,
+     vol. 35, no. 6, 2014, pp. 766–778., doi:10.1002/humu.22561.
+
+───────
+"""
+
+import argparse
+import logging
+import sys
+from pathlib import Path
+
+from ._upload_vcf import DEFAULT_REMOTE_URL
+from ._upload_vcf import HumanGenomeAssembly
+from ._upload_vcf import upload_vcf
+
+if __name__ == "__main__":
+    formatter = argparse.RawTextHelpFormatter
+
+    cli_args = sys.argv[1:]
+
+    parser = argparse.ArgumentParser(
+        description=__doc__,
+        add_help=True,
+        formatter_class=formatter,
+        epilog=r"Copyright © Clint Valentine 2024",
+    )
+
+    _ = parser.add_argument(
+        "--input",
+        required=True,
+        type=Path,
+        help="The path to the VCF to upload.",
+    )
+    _ = parser.add_argument(
+        "--email",
+        required=True,
+        type=str,
+        help="The email address to receive annotated variants at.",
+    )
+    _ = parser.add_argument(
+        "--assembly",
+        type=HumanGenomeAssembly,
+        default=HumanGenomeAssembly.hg38,
+        help="The human genome assembly of the VCF.\n(default: hg38)",
+    )
+    _ = parser.add_argument(
+        "--url",
+        type=str,
+        default=DEFAULT_REMOTE_URL,
+        help="The Seshat TP53 web server URL.\n(default: http://vps338341.ovh.net/batch_analysis)",
+    )
+    _ = parser.add_argument(
+        "--wait_for",
+        help="Seconds to wait for upload to occur before failure.\n(default: 5)",
+        default=5,
+    )
+    args = parser.parse_args(cli_args)
+
+    logging.basicConfig(datefmt="[%X]", level=logging.INFO)
+
+    upload_vcf(
+        vcf=args.input,
+        email=args.email,
+        assembly=args.assembly,
+        url=args.url,
+        wait_for=args.wait_for,
+    )
diff --git a/tp53/seshat/upload_vcf/_upload_vcf.py b/tp53/seshat/upload_vcf/_upload_vcf.py
@@ -1,18 +1,20 @@
+import logging
+import time
 from datetime import datetime
 from datetime import timedelta
 from enum import StrEnum
 from enum import auto
 from logging import Logger
-from logging import getLogger
 from pathlib import Path
 
+from chromedriver_py import binary_path
 from selenium import webdriver
 from selenium.webdriver.common.by import By
 from selenium.webdriver.remote.webdriver import WebDriver as RemoteWebDriver
 
-from ._exceptions import SeshatError
+from .._exceptions import SeshatError
 
-logger: Logger = getLogger("tp53.seshat")
+logger: Logger = logging.getLogger("tp53.seshat.upload_vcf")
 
 DEFAULT_REMOTE_URL: str = "http://vps338341.ovh.net/batch_analysis"
 """The default remote Seshat batch analysis URL."""
@@ -34,7 +36,7 @@ class HumanGenomeAssembly(StrEnum):
     """The human genome assembly GRCh37 (hg19)."""
 
 
-def seshat_upload_status(driver: RemoteWebDriver) -> str:
+def upload_status(driver: RemoteWebDriver) -> str:
     """Query the file uploading status and return its text representation."""
     modal = driver.find_element(By.XPATH, '//*[@id="uploading-status-text"]')
     inner = modal.get_attribute("innerText")
@@ -55,16 +57,18 @@ def upload_vcf(
 
     Args:
         vcf: The path to the VCF to upload.
-        email: The email address to receive Seshat TP53 variant annotations.
+        email: The email address to receive annotated variants at.
         assembly: The human genome assembly of the VCF.
         url: The Seshat TP53 web server URL.
-        wait_for: The total amount of time in seconds to wait for the upload occur before failure.
+        wait_for: Seconds to wait for upload to occur before failure.
     """
     vcf = str(Path(vcf).expanduser().absolute())
 
+    service = webdriver.ChromeService(executable_path=binary_path)
     options = webdriver.ChromeOptions()
     options.add_argument("headless")
-    driver = webdriver.Chrome(options=options)
+
+    driver = webdriver.Chrome(service=service, options=options)
 
     driver.get(url)
     driver.find_element(By.XPATH, f'//select[@id="reference"]/option[@value="{assembly}"]').click()
@@ -75,8 +79,9 @@ def upload_vcf(
 
     status: str = ""
     while (SUCCESS not in status) and datetime.now() < upload_start + timedelta(seconds=wait_for):
-        status = seshat_upload_status(driver)
+        status = upload_status(driver)
         logger.info(status)
+        time.sleep(0.1)
 
     driver.quit()
 

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-from tp53.seshat import HumanGenomeAssembly`
	`1`	`+from tp53.seshat.upload_vcf import HumanGenomeAssembly`
`2`	`2`
`3`	`3`
`4`	`4`	`def test_human_genome_assembly() -> None:`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1 @@`
	`1`	`+from ._find_in_gmail import find_in_gmail as find_in_gmail`
Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,2 @@`
	`1`	`+from ._upload_vcf import HumanGenomeAssembly as HumanGenomeAssembly`
	`2`	`+from ._upload_vcf import upload_vcf as upload_vcf`