attempting to reduce code bloat for find_clones (zktuong#329)

* how about this? * ok this implements the functions. got to think about how to adjust the loops to allow for the orphan vj clonotypes * ok got find_clones more or less working with the update. need to check generate_network to ensure that the overlap is computed properly * ok one more fix * Update test_tools_light.py * Update 3_dandelion_findingclones-10x_data.ipynb * Update _tools.py didn't realise pandas has a max number of zeroes. * minor wording change and bug fix * update notebooks
HongjianSun · Dec 12, 2023 · 6f4593d · 6f4593d
1 parent 8b2fca9
commit 6f4593d
Show file tree

Hide file tree

Showing 15 changed files with 2,113 additions and 2,097 deletions.
diff --git a/dandelion/preprocessing/_preprocessing.py b/dandelion/preprocessing/_preprocessing.py
@@ -2390,8 +2390,8 @@ def quantify_mutations(
             time=start,
             deep=(
                 "Updated Dandelion object: \n"
-                "   'data', contig-indexed clone table\n"
-                "   'metadata', cell-indexed clone table\n"
+                "   'data', contig-indexed AIRR table\n"
+                "   'metadata', cell-indexed observations table\n"
             ),
         )
     else:

diff --git a/dandelion/tools/_network.py b/dandelion/tools/_network.py
@@ -87,7 +87,9 @@ def generate_network(
     """
     regenerate = True
     if vdj_data.graph is not None:
-        if use_existing_graph:
+        if (min_size != 2) or (downsample is not None):
+            pass
+        elif use_existing_graph:
             start = logg.info(
                 "Generating network layout from pre-computed network"
             )
@@ -485,8 +487,8 @@ def generate_network(
         time=start,
         deep=(
             "Updated Dandelion object: \n"
-            "   'data', contig-indexed clone table\n"
-            "   'metadata', cell-indexed clone table\n"
+            "   'data', contig-indexed AIRR table\n"
+            "   'metadata', cell-indexed observations table\n"
             "   'layout', graph layout\n"
             "   'graph', network constructed from distance matrices of VDJ- and VJ- chains"
         ),

diff --git a/dandelion/tools/_tools.py b/dandelion/tools/_tools.py
diff --git a/dandelion/utilities/_core.py b/dandelion/utilities/_core.py
@@ -1128,7 +1128,7 @@ def write_h5ddl(
         if self.metadata is not None:
             metadata = self.metadata.copy()
             for col in metadata.columns:
-                if sys.version_info < (3, 9):
+                if pd.__version__ < "2.1.0":
                     weird = (
                         metadata[[col]].applymap(type)
                         != metadata[[col]].iloc[0].apply(type)

diff --git a/dandelion/utilities/_utilities.py b/dandelion/utilities/_utilities.py
@@ -21,6 +21,11 @@
 FALSES = ["F", "False", "false", "FALSE", False]
 HEAVYLONG = ["IGH", "TRB", "TRD"]
 LIGHTSHORT = ["IGK", "IGL", "TRA", "TRG"]
+VCALL = "v_call"
+JCALL = "j_call"
+VCALLG = "v_call_genotyped"
+JCALLG = "j_call_genotyped"
+STRIPALLELENUM = "[*][0-9][0-9]"
 
 # for compatibility with python>=3.10
 try:
@@ -1094,3 +1099,8 @@ def sum_col(vals):
         return np.nan
     else:
         return sum(vals)
+
+
+def check_same_celltype(clone_def1, clone_def2):
+    """Check if the first key is the same."""
+    return clone_def1.split("_", 1)[0] == clone_def2.split("_", 1)[0]
diff --git a/docs/notebooks/0_dandelion_primer.ipynb b/docs/notebooks/0_dandelion_primer.ipynb
diff --git a/docs/notebooks/1b_dandelion_noreannotation-10x_data.ipynb b/docs/notebooks/1b_dandelion_noreannotation-10x_data.ipynb