From d4a26c31722330048262e78d9a61b8c00bd2cd91 Mon Sep 17 00:00:00 2001
From: omesser <odedmesser@gmail.com>
Date: Thu, 13 Jul 2023 23:05:19 +0300
Subject: [PATCH 1/8] Use importlib package to dist resolution

---
 mlem/utils/module.py             | 28 ++++++++++++++++++----------
 setup.py                         |  5 ++++-
 tests/utils/test_module_tools.py | 11 +++++++----
 tests/utils/test_save.ipynb      | 32 +++++++++++++++++---------------
 4 files changed, 46 insertions(+), 30 deletions(-)

diff --git a/mlem/utils/module.py b/mlem/utils/module.py
index 4099cf7d..910c8272 100644
--- a/mlem/utils/module.py
+++ b/mlem/utils/module.py
@@ -8,6 +8,7 @@
 import sys
 import threading
 import warnings
+import importlib.metadata
 from collections import defaultdict
 from functools import lru_cache, wraps
 from pickle import PickleError
@@ -249,15 +250,22 @@ def get_module_version(mod: ModuleType) -> Optional[str]:
     :param mod: module object to use
     :return: version as `str` or `None` if version could not be determined
     """
-    for attr in "__version__", "VERSION":
-        if hasattr(mod, attr):
-            return str(getattr(mod, attr))
-    if mod.__file__ is None:
-        return None
-    for name in os.listdir(os.path.dirname(mod.__file__)):
-        m = re.match(re.escape(mod.__name__) + "-(.+)\\.dist-info", name)
-        if m:
-            return m.group(1)
+    # try using importlib package -> distro mapping and distro metadata
+    package_to_distros = packages_distributions()
+    try:
+        if mod.__name__ in package_to_distros:
+            for distro_name in package_to_distros[mod.__name__]:
+                distro = importlib.metadata.distribution(distro_name)
+                return distro.version
+    except importlib.metadata.PackageNotFoundError:
+        pass
+
+    # if there's a package-file, try to get it from there
+    if mod.__file__ is not None:
+        for name in os.listdir(os.path.dirname(mod.__file__)):
+            m = re.match(re.escape(mod.__name__) + "-(.+)\\.dist-info", name)
+            if m:
+                return m.group(1)
     return None
 
 
@@ -502,7 +510,7 @@ def save_type_with_classvars(pickler: "RequirementAnalyzer", obj):
 
 class RequirementAnalyzer(dill.Pickler):
     """Special pickler implementation that collects requirements while pickling
-    (and not pickling actualy)"""
+    (and not actually pickling)"""
 
     ignoring = (
         "dill",
diff --git a/setup.py b/setup.py
index 526b3356..e1e3a584 100644
--- a/setup.py
+++ b/setup.py
@@ -57,6 +57,9 @@
     "jupyter",
     "nbconvert",
     "nbloader",
+    # We're using regex to test requirement version extraction
+    # edge case, see: https://github.com/iterative/mlem/issues/688
+    'regex==2023.6.3',
 ]
 
 extras = {
@@ -75,7 +78,7 @@
     "lightgbm": ["lightgbm"],
     "fastapi": ["uvicorn", "fastapi"],
     "prometheus": ["prometheus-fastapi-instrumentator"],
-    "streamlit": ["uvicorn", "fastapi", "streamlit", "streamlit_pydantic"],
+    "streamlit": ["uvicorn", "fastapi", "streamlit>=1.14.0", "streamlit_pydantic"],
     "sagemaker": ["docker", "boto3", "sagemaker"],
     "torch": ["torch"],
     "tensorflow": ["tensorflow"],
diff --git a/tests/utils/test_module_tools.py b/tests/utils/test_module_tools.py
index f971f493..b4f0b4d7 100644
--- a/tests/utils/test_module_tools.py
+++ b/tests/utils/test_module_tools.py
@@ -1,6 +1,7 @@
 import os
 import subprocess
 from typing import ClassVar
+import regex
 
 import nbformat
 import numpy
@@ -220,11 +221,11 @@ def test_get_requirements_notebook():
 
     loaded_notebook = Notebook(TEST_NOTEBOOK_NAME)
 
-    kek = loaded_notebook.run_all()
-    res = kek.ns["res"]
+    notebook = loaded_notebook.run_all()
+    res = notebook.ns["res"]
 
     assert isinstance(res, Requirements)
-    assert res.modules == ["numpy"]
+    assert sorted(res.modules) == sorted(["regex", "numpy"])
 
 
 def _run_jup(command):
@@ -249,7 +250,9 @@ def test_get_requirements_notebook_run():
     with open(TEST_NOTEBOOK_NAME, encoding="utf8") as f:
         nb = nbformat.read(f, as_version=4)
 
-    assert nb["cells"][1]["outputs"][0].text.strip() == get_module_repr(numpy)
+    res = nb["cells"][1]["outputs"][0].text.strip()
+    assert get_module_repr(numpy) in res
+    assert get_module_repr(regex) in res
 
 
 # Copyright 2019 Zyfra
diff --git a/tests/utils/test_save.ipynb b/tests/utils/test_save.ipynb
index 68ccf781..b6c97dc8 100644
--- a/tests/utils/test_save.ipynb
+++ b/tests/utils/test_save.ipynb
@@ -5,10 +5,10 @@
    "execution_count": 1,
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-02-13T14:11:29.261665Z",
-     "iopub.status.busy": "2023-02-13T14:11:29.261394Z",
-     "iopub.status.idle": "2023-02-13T14:11:29.267566Z",
-     "shell.execute_reply": "2023-02-13T14:11:29.266734Z"
+     "iopub.execute_input": "2023-07-13T20:04:23.425791Z",
+     "iopub.status.busy": "2023-07-13T20:04:23.425682Z",
+     "iopub.status.idle": "2023-07-13T20:04:23.436620Z",
+     "shell.execute_reply": "2023-07-13T20:04:23.436325Z"
     },
     "pycharm": {
      "name": "#%%\n"
@@ -17,8 +17,10 @@
    "outputs": [],
    "source": [
     "import numpy as np\n",
+    "import regex\n",
     "\n",
-    "def func(data):\n",
+    "def use_modules():\n",
+    "    regex.split(' ', 'brown fox')\n",
     "    return bool(np.all([True]))\n"
    ]
   },
@@ -27,13 +29,12 @@
    "execution_count": 2,
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-02-13T14:11:29.270506Z",
-     "iopub.status.busy": "2023-02-13T14:11:29.270233Z",
-     "iopub.status.idle": "2023-02-13T14:11:29.969468Z",
-     "shell.execute_reply": "2023-02-13T14:11:29.968705Z"
+     "iopub.execute_input": "2023-07-13T20:04:23.438339Z",
+     "iopub.status.busy": "2023-07-13T20:04:23.438249Z",
+     "iopub.status.idle": "2023-07-13T20:04:24.532574Z",
+     "shell.execute_reply": "2023-07-13T20:04:24.532277Z"
     },
     "pycharm": {
-     "is_executing": true,
      "name": "#%%\n"
     }
    },
@@ -42,16 +43,17 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "numpy==1.22.4\n"
+      "regex==2023.6.3\n",
+      "numpy==1.25.1\n"
      ]
     }
    ],
    "source": [
     "from mlem.utils.module import get_object_requirements\n",
     "\n",
-    "res = get_object_requirements(func)\n",
+    "res = get_object_requirements(use_modules)\n",
     "\n",
-    "print(\" \".join(res.to_pip()))"
+    "print(\"\\n\".join(res.to_pip()))"
    ]
   }
  ],
@@ -71,9 +73,9 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.9.13"
+   "version": "3.10.9"
   }
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}
\ No newline at end of file

From 2572a1517eea60cbd3e3c1c2a58010d70d7a2e9c Mon Sep 17 00:00:00 2001
From: omesser <odedmesser@gmail.com>
Date: Thu, 13 Jul 2023 23:11:07 +0300
Subject: [PATCH 2/8] Slightly less verbose

---
 mlem/utils/module.py | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/mlem/utils/module.py b/mlem/utils/module.py
index 910c8272..e009b813 100644
--- a/mlem/utils/module.py
+++ b/mlem/utils/module.py
@@ -8,7 +8,7 @@
 import sys
 import threading
 import warnings
-import importlib.metadata
+from importlib.metadata import distribution, PackageNotFoundError
 from collections import defaultdict
 from functools import lru_cache, wraps
 from pickle import PickleError
@@ -255,9 +255,9 @@ def get_module_version(mod: ModuleType) -> Optional[str]:
     try:
         if mod.__name__ in package_to_distros:
             for distro_name in package_to_distros[mod.__name__]:
-                distro = importlib.metadata.distribution(distro_name)
+                distro = distribution(distro_name)
                 return distro.version
-    except importlib.metadata.PackageNotFoundError:
+    except PackageNotFoundError:
         pass
 
     # if there's a package-file, try to get it from there

From 08b8152ef61167149b79a00474489774a11695d0 Mon Sep 17 00:00:00 2001
From: omesser <odedmesser@gmail.com>
Date: Thu, 13 Jul 2023 23:19:54 +0300
Subject: [PATCH 3/8] fixes, linting and such

---
 mlem/utils/module.py             | 2 +-
 setup.py                         | 4 ++--
 tests/utils/test_module_tools.py | 2 +-
 tests/utils/test_save.ipynb      | 2 +-
 4 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/mlem/utils/module.py b/mlem/utils/module.py
index e009b813..a6163c55 100644
--- a/mlem/utils/module.py
+++ b/mlem/utils/module.py
@@ -8,9 +8,9 @@
 import sys
 import threading
 import warnings
-from importlib.metadata import distribution, PackageNotFoundError
 from collections import defaultdict
 from functools import lru_cache, wraps
+from importlib.metadata import PackageNotFoundError, distribution
 from pickle import PickleError
 from types import FunctionType, LambdaType, MethodType, ModuleType
 from typing import Dict, List, Optional, Set, Union
diff --git a/setup.py b/setup.py
index e1e3a584..d999d91a 100644
--- a/setup.py
+++ b/setup.py
@@ -59,7 +59,7 @@
     "nbloader",
     # We're using regex to test requirement version extraction
     # edge case, see: https://github.com/iterative/mlem/issues/688
-    'regex==2023.6.3',
+    "regex==2023.6.3",
 ]
 
 extras = {
@@ -78,7 +78,7 @@
     "lightgbm": ["lightgbm"],
     "fastapi": ["uvicorn", "fastapi"],
     "prometheus": ["prometheus-fastapi-instrumentator"],
-    "streamlit": ["uvicorn", "fastapi", "streamlit>=1.14.0", "streamlit_pydantic"],
+    "streamlit": ["uvicorn", "fastapi", "streamlit", "streamlit_pydantic"],
     "sagemaker": ["docker", "boto3", "sagemaker"],
     "torch": ["torch"],
     "tensorflow": ["tensorflow"],
diff --git a/tests/utils/test_module_tools.py b/tests/utils/test_module_tools.py
index b4f0b4d7..02253434 100644
--- a/tests/utils/test_module_tools.py
+++ b/tests/utils/test_module_tools.py
@@ -1,11 +1,11 @@
 import os
 import subprocess
 from typing import ClassVar
-import regex
 
 import nbformat
 import numpy
 import pytest
+import regex
 from pydantic import BaseModel
 
 from mlem.core.requirements import Requirements
diff --git a/tests/utils/test_save.ipynb b/tests/utils/test_save.ipynb
index b6c97dc8..2ef30103 100644
--- a/tests/utils/test_save.ipynb
+++ b/tests/utils/test_save.ipynb
@@ -78,4 +78,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
\ No newline at end of file
+}

From 41f36fc7823ab538f1662b5e5d882e215c71a00b Mon Sep 17 00:00:00 2001
From: omesser <odedmesser@gmail.com>
Date: Thu, 13 Jul 2023 23:40:36 +0300
Subject: [PATCH 4/8] Take free-space.sh from nuclio

---
 tests/utils/test_save.ipynb | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tests/utils/test_save.ipynb b/tests/utils/test_save.ipynb
index 2ef30103..b6c97dc8 100644
--- a/tests/utils/test_save.ipynb
+++ b/tests/utils/test_save.ipynb
@@ -78,4 +78,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
+}
\ No newline at end of file

From 0266c814e92d95392028e43e0cd0ab352a0b01db Mon Sep 17 00:00:00 2001
From: omesser <odedmesser@gmail.com>
Date: Thu, 13 Jul 2023 23:44:10 +0300
Subject: [PATCH 5/8] Fixing ipynb lint again. shakes-fist-at-pycharm

---
 tests/utils/test_save.ipynb | 32 +++++++++++++++-----------------
 1 file changed, 15 insertions(+), 17 deletions(-)

diff --git a/tests/utils/test_save.ipynb b/tests/utils/test_save.ipynb
index b6c97dc8..68ccf781 100644
--- a/tests/utils/test_save.ipynb
+++ b/tests/utils/test_save.ipynb
@@ -5,10 +5,10 @@
    "execution_count": 1,
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-07-13T20:04:23.425791Z",
-     "iopub.status.busy": "2023-07-13T20:04:23.425682Z",
-     "iopub.status.idle": "2023-07-13T20:04:23.436620Z",
-     "shell.execute_reply": "2023-07-13T20:04:23.436325Z"
+     "iopub.execute_input": "2023-02-13T14:11:29.261665Z",
+     "iopub.status.busy": "2023-02-13T14:11:29.261394Z",
+     "iopub.status.idle": "2023-02-13T14:11:29.267566Z",
+     "shell.execute_reply": "2023-02-13T14:11:29.266734Z"
     },
     "pycharm": {
      "name": "#%%\n"
@@ -17,10 +17,8 @@
    "outputs": [],
    "source": [
     "import numpy as np\n",
-    "import regex\n",
     "\n",
-    "def use_modules():\n",
-    "    regex.split(' ', 'brown fox')\n",
+    "def func(data):\n",
     "    return bool(np.all([True]))\n"
    ]
   },
@@ -29,12 +27,13 @@
    "execution_count": 2,
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-07-13T20:04:23.438339Z",
-     "iopub.status.busy": "2023-07-13T20:04:23.438249Z",
-     "iopub.status.idle": "2023-07-13T20:04:24.532574Z",
-     "shell.execute_reply": "2023-07-13T20:04:24.532277Z"
+     "iopub.execute_input": "2023-02-13T14:11:29.270506Z",
+     "iopub.status.busy": "2023-02-13T14:11:29.270233Z",
+     "iopub.status.idle": "2023-02-13T14:11:29.969468Z",
+     "shell.execute_reply": "2023-02-13T14:11:29.968705Z"
     },
     "pycharm": {
+     "is_executing": true,
      "name": "#%%\n"
     }
    },
@@ -43,17 +42,16 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "regex==2023.6.3\n",
-      "numpy==1.25.1\n"
+      "numpy==1.22.4\n"
      ]
     }
    ],
    "source": [
     "from mlem.utils.module import get_object_requirements\n",
     "\n",
-    "res = get_object_requirements(use_modules)\n",
+    "res = get_object_requirements(func)\n",
     "\n",
-    "print(\"\\n\".join(res.to_pip()))"
+    "print(\" \".join(res.to_pip()))"
    ]
   }
  ],
@@ -73,9 +71,9 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.10.9"
+   "version": "3.9.13"
   }
  },
  "nbformat": 4,
  "nbformat_minor": 0
-}
\ No newline at end of file
+}

From ea167b55e8bdc76b52dc6fe93a6b95c511108e37 Mon Sep 17 00:00:00 2001
From: omesser <odedmesser@gmail.com>
Date: Mon, 31 Jul 2023 12:49:19 +0300
Subject: [PATCH 6/8] typo fix

---
 mlem/core/hooks.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mlem/core/hooks.py b/mlem/core/hooks.py
index 769cbcf8..c43af595 100644
--- a/mlem/core/hooks.py
+++ b/mlem/core/hooks.py
@@ -67,7 +67,7 @@ def __init_subclass__(cls, *args, **kwargs):
                 )
         else:
             logger.debug(
-                "Not registerting %s to any Analyzer because it's an abstract class",
+                "Not registering %s to any Analyzer because it's an abstract class",
                 cls.__name__,
             )
         super(Hook, cls).__init_subclass__(*args, **kwargs)

From 2059f214dfebe8a373d9a8c0eb925df6b45af87b Mon Sep 17 00:00:00 2001
From: omesser <odedmesser@gmail.com>
Date: Mon, 31 Jul 2023 13:04:25 +0300
Subject: [PATCH 7/8] fixing test_save.ipynb

---
 tests/utils/test_save.ipynb | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)

diff --git a/tests/utils/test_save.ipynb b/tests/utils/test_save.ipynb
index 68ccf781..275713f5 100644
--- a/tests/utils/test_save.ipynb
+++ b/tests/utils/test_save.ipynb
@@ -5,10 +5,10 @@
    "execution_count": 1,
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-02-13T14:11:29.261665Z",
-     "iopub.status.busy": "2023-02-13T14:11:29.261394Z",
-     "iopub.status.idle": "2023-02-13T14:11:29.267566Z",
-     "shell.execute_reply": "2023-02-13T14:11:29.266734Z"
+     "iopub.execute_input": "2023-07-31T10:02:34.099802Z",
+     "iopub.status.busy": "2023-07-31T10:02:34.099406Z",
+     "iopub.status.idle": "2023-07-31T10:02:34.115259Z",
+     "shell.execute_reply": "2023-07-31T10:02:34.114688Z"
     },
     "pycharm": {
      "name": "#%%\n"
@@ -17,8 +17,11 @@
    "outputs": [],
    "source": [
     "import numpy as np\n",
+    "import regex\n",
     "\n",
     "def func(data):\n",
+    "    # using regex, just so it's listed and its version inferred correctly\n",
+    "    regex.search(r'(ab)', 'abcdef')\n",
     "    return bool(np.all([True]))\n"
    ]
   },
@@ -27,10 +30,10 @@
    "execution_count": 2,
    "metadata": {
     "execution": {
-     "iopub.execute_input": "2023-02-13T14:11:29.270506Z",
-     "iopub.status.busy": "2023-02-13T14:11:29.270233Z",
-     "iopub.status.idle": "2023-02-13T14:11:29.969468Z",
-     "shell.execute_reply": "2023-02-13T14:11:29.968705Z"
+     "iopub.execute_input": "2023-07-31T10:02:34.118394Z",
+     "iopub.status.busy": "2023-07-31T10:02:34.118192Z",
+     "iopub.status.idle": "2023-07-31T10:02:35.897692Z",
+     "shell.execute_reply": "2023-07-31T10:02:35.897269Z"
     },
     "pycharm": {
      "is_executing": true,
@@ -42,7 +45,7 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "numpy==1.22.4\n"
+      "numpy==1.25.1 regex==2023.6.3\n"
      ]
     }
    ],
@@ -71,7 +74,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.9.13"
+   "version": "3.10.9"
   }
  },
  "nbformat": 4,

From 02232ca10e06d81cb070acf24a57bf6c1cfc7aa4 Mon Sep 17 00:00:00 2001
From: omesser <odedmesser@gmail.com>
Date: Wed, 23 Aug 2023 16:21:26 +0300
Subject: [PATCH 8/8] Adding pragma no branch

---
 mlem/utils/module.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mlem/utils/module.py b/mlem/utils/module.py
index a6163c55..260c8231 100644
--- a/mlem/utils/module.py
+++ b/mlem/utils/module.py
@@ -261,7 +261,7 @@ def get_module_version(mod: ModuleType) -> Optional[str]:
         pass
 
     # if there's a package-file, try to get it from there
-    if mod.__file__ is not None:
+    if mod.__file__ is not None:  # pragma: no branch
         for name in os.listdir(os.path.dirname(mod.__file__)):
             m = re.match(re.escape(mod.__name__) + "-(.+)\\.dist-info", name)
             if m: