Merge pull request #356 from RavirajBaraiya/pip_zingg

zingg pip package artifacts modification

python/MANIFEST.in

@@ -0,0 +1,14 @@
+global-exclude *.py[cod] __pycache__ .DS_Store *.pyc
+recursive-include deps/jars *.jar
+graft deps/scripts
+recursive-include deps/scripts zingg.sh
+recursive-include deps/licenses LICENSE
+recursive-include deps/examples *
+recursive-include deps/models *
+recursive-include deps/phases *
+recursive-include deps/config *
+recursive-include lib *.zip
+include *.py
+include LICENSE
+include README.md
+include requirements.txt

python/PKG-INFO

@@ -0,0 +1,28 @@
+Metadata-Version: 2.1
+Name: zingg
+Version: 0.3.4
+Summary: Zingg.ai Entity Resolution
+Home-page: www.zingg.ai
+Author: Zingg.AI
+Author-email: [email protected]
+License: UNKNOWN
+Description: ## About Zingg
+
+        Zingg is an ML based entity resolution framework. The Python Package is used for building training data, training Zingg models and running the matching and linking processes.
+        [Zingg.AI](https://www.zingg.ai) 
+
+        # requirement
+        python 3.6+; spark 3.1.2
+
+        # Installation
+
+        ```
+        pip install zingg
+        ``` 
+
+
+Platform: UNKNOWN
+Classifier: Programming Language :: Python :: 3
+Classifier: License :: OSI Approved :: GNU Affero General Public License v3
+Classifier: Operating System :: OS Independent
+Description-Content-Type: text/markdown 

python/install.py

@@ -0,0 +1,73 @@
+import os
+import re
+import tarfile
+import traceback
+import urllib.request
+from shutil import rmtree
+
+def checked_package_name(zingg_version):
+    """
+    Check the generated package name.
+    """
+    return "%s" % (zingg_version)
+
+def checked_versions(zingg_version):
+    """
+    Check the valid combinations of supported versions in zingg distributions.
+
+    Parameters
+    ----------
+    zingg_version : str
+        zingg version. It should be X.X.X such as '3.0.0' or zingg-3.0.0.
+    """
+    if re.match("^[0-9]+\\.[0-9]+\\.[0-9]+$", zingg_version):
+        zingg_version = "zingg-%s" % zingg_version
+    if not zingg_version.startswith("zingg-"):
+        raise RuntimeError(
+            "zingg version should start with 'zingg-' prefix; however, " "got %s" % zingg_version
+        )
+
+    return zingg_version
+
+def install_zingg(dest, zingg_version):
+    """
+    Installs zingg that corresponds to the given Hadoop version in the current
+    library directory.
+
+    Parameters
+    ----------
+    dest : str
+        The location to download and install the zingg.
+    zingg_version : str
+        zingg version. It should be zingg-X.X.X form.
+    """
+
+    package_name = checked_package_name(zingg_version)
+    package_local_path = os.path.join(dest, "%s.tgz" % package_name)
+
+    os.makedirs(dest, exist_ok=True)
+
+    tar = None
+
+    try:
+
+        print("Installing to %s" % dest)
+        tar = tarfile.open(package_local_path, "r:gz")
+        for member in tar.getmembers():
+            if member.name == package_name:
+                # Skip the root directory.
+                continue
+            member.name = os.path.relpath(member.name, package_name + os.path.sep)
+            tar.extract(member, dest)
+        return
+    except Exception:
+        print("Failed to open TAR")
+        traceback.print_exc()
+        rmtree(dest, ignore_errors=True)
+    finally:
+        if tar is not None:
+            tar.close()
+        if os.path.exists(package_local_path):
+            os.remove(package_local_path)
+
+    raise IOError("Unable to run INSTALL")

python/phases/assessModel.py

@@ -1,4 +1,4 @@
-from python.api.zingg.zingg import *
+from zingg import *
 import pandas as pd
 import seaborn as sn
 import matplotlib.pyplot as plt

python/phases/exportModel.py

@@ -1,5 +1,5 @@
 
-from python.api.zingg.zingg import *
+from zingg import *
 import sys
 import argparse
 import os

python/setup.cfg

@@ -0,0 +1,5 @@
+[bdist_wheel]
+universal = 1
+
+[metadata]
+description-file = README.md