fix: Restore python rules changes triggering reruns. (GH: #3213)

royjacobson · royjacobson · commit ea0782f9ec20 · 2025-03-26T15:14:33.000+02:00
diff --git a/src/snakemake/persistence.py b/src/snakemake/persistence.py
@@ -553,11 +553,13 @@ def _b64id(self, s):
 
     @lru_cache()
     def _code(self, rule):
-        # We only consider shell commands for now.
-        # Plain python code rules are hard to capture because the pickling of the code
-        # can change with different python versions.
         # Scripts and notebooks are triggered by changes in the script mtime.
-        return rule.shellcmd if rule.shellcmd is not None else None
+        # Changes to python and shell rules are triggered by changes in the plain text.
+        if rule.shellcmd is not None:
+            return rule.shellcmd
+        if rule.run_func_src is not None:
+            return rule.run_func_src
+        return None
 
     @lru_cache()
     def _conda_env(self, job):
diff --git a/src/snakemake/rules.py b/src/snakemake/rules.py
@@ -107,6 +107,7 @@ def __init__(self, name, workflow, lineno=None, snakefile=None):
         self._lineno = lineno
         self._snakefile = snakefile
         self.run_func = None
+        self.run_func_src = None
         self.shellcmd = None
         self.script = None
         self.notebook = None
diff --git a/src/snakemake/workflow.py b/src/snakemake/workflow.py
@@ -10,6 +10,8 @@
 import subprocess
 import sys
 import platform
+import dis
+import linecache
 from collections import OrderedDict, namedtuple
 from collections.abc import Mapping
 from itertools import filterfalse, chain
@@ -1712,6 +1714,26 @@ def ruleorder(self, *rulenames):
     def localrules(self, *rulenames):
         self._localrules.update(rulenames)
 
+    def get_rule_source(self, func):
+        # This can't use `inspect` because the functions are compiled into intermediate python code
+        # in parser.py and that intermediate source is not available anymore (or desirable).
+        # Instead, we're using dis to retrieve the line numbers of the function in the intermediate
+        # code and then map it to the original file using `self.linemaps`.
+        sourcefile = func.__code__.co_filename
+        line_numbers = []
+        linemap = self.linemaps[sourcefile]
+        for func_offset, line in dis.findlinestarts(func.__code__):
+            # The first instruction in the compiled function is RESUME, which
+            # with snakemake is mapped to the 'rule: ' line and is not considered
+            # part of the rule source.
+            if func_offset == 0:
+                continue
+            if line in linemap:
+                line_numbers.append(linemap[line])
+        return "".join(
+            [linecache.getline(sourcefile, lineno) for lineno in sorted(line_numbers)]
+        )
+
     def rule(self, name=None, lineno=None, snakefile=None, checkpoint=False):
         # choose a name for an unnamed rule
         if name is None:
@@ -1930,6 +1952,8 @@ def check_may_use_software_deployment(method):
                 rule.name = ruleinfo.name
             rule.docstring = ruleinfo.docstring
             rule.run_func = ruleinfo.func
+            if rule.run_func is not None:
+                rule.run_func_src = self.get_rule_source(rule.run_func)
             rule.shellcmd = ruleinfo.shellcmd
             rule.script = ruleinfo.script
             rule.notebook = ruleinfo.notebook
