diff --git a/src/dynapyt/run_analysis.py b/src/dynapyt/run_analysis.py
index e63db94..9e596a3 100644
--- a/src/dynapyt/run_analysis.py
+++ b/src/dynapyt/run_analysis.py
@@ -3,20 +3,31 @@
 import importlib
 from os.path import abspath
 from tempfile import gettempdir
-from shutil import rmtree
 import sys
 from pathlib import Path
 from . import runtime as _rt
 
 
 def run_analysis(
-    entry: str, analyses: List[str], name: str = None, coverage: bool = False
+    entry: str,
+    analyses: List[str],
+    name: str = None,
+    coverage: bool = False,
+    coverage_dir: str = None,
 ):
-    coverage_dir = Path(gettempdir()) / "dynapyt_coverage"
+    global _rt
+    _rt = importlib.reload(_rt)
+
     if coverage:
-        coverage_dir.mkdir(exist_ok=True)
+        if coverage_dir is None:
+            coverage_path = Path(gettempdir()) / "dynapyt_coverage"
+            coverage_path.mkdir(exist_ok=True)
+        else:
+            coverage_path = Path(coverage)
+            coverage_path.mkdir(exist_ok=True)
+        _rt.set_coverage(coverage_path)
     else:
-        rmtree(str(coverage_dir), ignore_errors=True)
+        _rt.set_coverage(None)
 
     analyses_file = Path(gettempdir()) / "dynapyt_analyses.txt"
     if analyses_file.exists():
@@ -38,6 +49,8 @@ def run_analysis(
         globals_dict["__file__"] = entry_full_path
         exec(open(entry_full_path).read(), globals_dict)
     else:
+        if importlib.util.find_spec(entry) is None:
+            raise ValueError(f"Could not find entry {entry}")
         importlib.import_module(entry)
     _rt.end_execution()
 
diff --git a/src/dynapyt/runtime.py b/src/dynapyt/runtime.py
index 89d87c9..844391d 100644
--- a/src/dynapyt/runtime.py
+++ b/src/dynapyt/runtime.py
@@ -15,20 +15,19 @@
 
 analyses = None
 covered = None
+coverage_path = None
 current_file = None
 end_execution_called = False
 
 
 def end_execution():
-    global covered, end_execution_called
+    global covered, coverage_path, end_execution_called
     if end_execution_called:
         return
     end_execution_called = True
     call_if_exists("end_execution")
     if covered is not None:
-        coverage_file = (
-            Path(tempfile.gettempdir()) / "dynapyt_coverage" / "covered.jsonl"
-        )
+        coverage_file = coverage_path / "covered.jsonl"
         with FileLock(f"{str(coverage_file)}.lock"):
             if coverage_file.exists():
                 existing_coverage = {}
@@ -59,15 +58,25 @@ def end_execution():
 
 def set_analysis(new_analyses: List[Any]):
     global analyses, covered
-    if analyses is None:
-        analyses = []
-        coverage_dir = Path(tempfile.gettempdir()) / "dynapyt_coverage"
-        if coverage_dir.exists():
-            covered = {}
-        signal.signal(signal.SIGINT, end_execution)
-        signal.signal(signal.SIGTERM, end_execution)
-        atexit.register(end_execution)
-        analyses = load_analyses(new_analyses)
+    analyses = []
+    signal.signal(signal.SIGINT, end_execution)
+    signal.signal(signal.SIGTERM, end_execution)
+    atexit.register(end_execution)
+    analyses = load_analyses(new_analyses)
+
+
+def set_coverage(coverage_path: Path):
+    if coverage_path is not None:
+        global covered
+        covered = {}
+        coverage_path.mkdir(exist_ok=True)
+        coverage_file = coverage_path / "covered.jsonl"
+        if coverage_file.exists():
+            with open(str(coverage_file), "r") as f:
+                content = f.read().splitlines()
+            for c in content:
+                tmp = json.loads(c)
+                covered.update(tmp)
 
 
 def filtered(func, f, args):
diff --git a/tests/run_single_test.py b/tests/run_single_test.py
index 1e3ed08..4baceee 100644
--- a/tests/run_single_test.py
+++ b/tests/run_single_test.py
@@ -68,22 +68,19 @@ def test_runner(directory_pair: Tuple[str, str], capsys):
         instrument_file(join(abs_dir, "__init__.py"), selected_hooks)
 
     # analyze
-    # class_ = getattr(module, "TestAnalysis")
-    analysis_instances = [class_[1]() for class_ in analysis_classes]
-    rt.analyses = None
-    rt.set_analysis(analysis_instances)
     captured = capsys.readouterr()  # clear stdout
     # print(f"Before analysis: {captured.out}")  # for debugging purposes
-    for analysis_instance in analysis_instances:
-        if hasattr(analysis_instance, "begin_execution"):
-            analysis_instance.begin_execution()
     if run_as_file:
         run_analysis(program_file, [f"{module_prefix}.analysis.TestAnalysis"])
     else:
-        import_module(f"{module_prefix}.program")
-    for analysis_instance in analysis_instances:
-        if hasattr(analysis_instance, "end_execution"):
-            analysis_instance.end_execution()
+        run_analysis(
+            f"{module_prefix}.program",
+            [
+                f"{module_prefix}.analysis.{ac[0]}"
+                for ac in analysis_classes
+                if not ac[0].endswith("BaseAnalysis")
+            ],
+        )
 
     # check output
     expected_file = join(abs_dir, "expected.txt")