From 36c9e217f7eb49f64c014cf7b2414d899bc02259 Mon Sep 17 00:00:00 2001
From: Jayoung Ryu <jayoung_ryu@g.harvard.edu>
Date: Tue, 14 Nov 2023 18:11:10 -0500
Subject: [PATCH] adding files for editing preference profiling

---
 bean/plotting/utils.py | 69 ++++++++++++++++++++++++++++++++++++++++++
 bin/bean-profile       | 34 +++++++++++++++++++++
 2 files changed, 103 insertions(+)
 create mode 100644 bean/plotting/utils.py
 create mode 100644 bin/bean-profile

diff --git a/bean/plotting/utils.py b/bean/plotting/utils.py
new file mode 100644
index 0000000..1233cfb
--- /dev/null
+++ b/bean/plotting/utils.py
@@ -0,0 +1,69 @@
+import argparse
+
+
+def parse_args():
+    print("  \n~~~BEAN Profile~~~")
+    print("-Profile editing patterns of your editor-")
+    print(
+        r"""
+    _ _                     __ _ _     
+  /  \ '\     _ __ _ _ ___ / _(_) |___ 
+  |   \  \   | '_ \ '_/ _ \  _| | / -_)
+   \   \  |  | .__/_| \___/_| |_|_\___|
+    `.__|/   |_|                       
+    """
+    )
+    parser = argparse.ArgumentParser()
+    parser.add_argument(
+        "bdata_path", help="Path to the ReporterScreen object to run QC on", type=str
+    )
+    parser.add_argument(
+        "-o",
+        "--output-prefix",
+        help="Output prefix of editing pattern report (prefix.html, prefix.ipynb). If not provided, base name of `bdata_path` is used.",
+        type=str,
+    )
+    parser.add_argument(
+        "--replicate-col",
+        help="Column name in `bdata.samples` that describes replicate ID.",
+        type=str,
+        default="rep",
+    )
+    parser.add_argument(
+        "--condition-col",
+        help="Column name in `bdata.samples` that describes experimental condition. (sorting bin, time, etc.)",
+        type=str,
+        default="bin",
+    )
+    parser.add_argument(
+        "--pam-col",
+        help="Column name describing PAM of each gRNA in `bdata.guides`.",
+        type=str,
+        default=None,
+    )
+    parser.add_argument(
+        "--control-condition",
+        help="Control condition where editing preference would be profiled at. Pre-filters data where `bdata.samples[condition_col] == control_condition`.",
+        type=str,
+        default="bulk",
+    )
+    parser.add_argument(
+        "-w",
+        "--window-length",
+        help="Window length of editing window of maximal editing efficiency to be identified. This window is used to quantify context specificity within the window.",
+        type=int,
+        default=6,
+    )
+
+    args = parser.parse_args()
+    if args.output_prefix is None:
+        args.output_prefix = f"{args.bdata_path.rsplit('.h5ad', 1)[0]}"
+    return args
+
+
+def check_args(args):
+    if args.window_length < 1:
+        raise ValueError(f"window_length {args.window_length} is too small.")
+    if args.window_length > 20:
+        raise ValueError(f"window_length {args.window_length} is too large.")
+    return args
diff --git a/bin/bean-profile b/bin/bean-profile
new file mode 100644
index 0000000..8b48daa
--- /dev/null
+++ b/bin/bean-profile
@@ -0,0 +1,34 @@
+#!/usr/bin/env python
+import os
+import papermill as pm
+import bean as be
+from bean.plotting.utils import parse_args, check_args
+
+
+def main():
+    args = parse_args()
+    args = check_args(args)
+    os.system(
+        "python -m ipykernel install --user --name bean_python3 --display-name bean_python3"
+    )
+    pm.execute_notebook(
+        f"{os.path.dirname(be.__file__)}/../notebooks/profile_editing_preference.ipynb",
+        f"{args.output_prefix}.ipynb",
+        parameters=dict(
+            bdata_path=args.bdata_path,
+            output_prefix=args.output_prefix,
+            replicate_col=args.replicate_col,
+            condition_col=args.condition_col,
+            control_condition=args.control_condition,
+            max_editing_window_length=args.window_length,
+            pam_col=args.pam_col,
+        ),
+        kernel_name="bean_python3",
+    )
+    os.system(
+        f"jupyter nbconvert --to html {args.output_prefix}_editing_preference.ipynb"
+    )
+
+
+if __name__ == "__main__":
+    main()