From e7801268be189c7d09eb642c37c3be6eb06ed9fb Mon Sep 17 00:00:00 2001
From: "Samir J. Patel" <samirpa@us.ibm.com>
Date: Tue, 5 Jul 2016 11:32:13 -0400
Subject: [PATCH 1/7] [document-conversion] Adds index document API

This commit adds the index_document API for the document
conversion service for the python SDK. It also includes
updated examples for using the index_document API.
---
 examples/document_conversion_v1.py            | 76 ++++++++++++++++++-
 .../document_conversion_v1.py                 | 13 ++++
 2 files changed, 88 insertions(+), 1 deletion(-)

diff --git a/examples/document_conversion_v1.py b/examples/document_conversion_v1.py
index 536ac0db0..8c7c17836 100644
--- a/examples/document_conversion_v1.py
+++ b/examples/document_conversion_v1.py
@@ -13,9 +13,83 @@
 with open(join(dirname(__file__), '../resources/example.html'), 'r') as document:
     config = {'conversion_target': DocumentConversionV1.NORMALIZED_HTML}
     print(document_conversion.convert_document(document=document, config=config, media_type='text/html')
-          .content.decode('utf-8'))
+          .content)
 
 # Example with JSON
 with open(join(dirname(__file__), '../resources/example.html'), 'r') as document:
     config['conversion_target'] = DocumentConversionV1.ANSWER_UNITS
     print(json.dumps(document_conversion.convert_document(document=document, config=config), indent=2))
+
+# Examples of index_document API
+print("########## Example of a dry run of index_document with only a document ##########")
+with open(join(dirname(__file__), '../resources/example.html'), 'r') as document:
+    config = {
+        'retrieve_and_rank': {
+            'dry_run':'true'
+        }
+    }
+    print(json.dumps(document_conversion.index_document(config=config, document=document), indent=2))
+
+print("########## Example of a dry run of index_document with only metadata ##########")
+config = {
+    'retrieve_and_rank': {
+        'dry_run':'true'
+    }
+}
+metadata = {
+    'metadata': [
+        {'name':'id', 'value':'12345'}
+    ]
+}
+print(json.dumps(document_conversion.index_document(config=config, metadata=metadata), indent=2))
+
+print("########## Example of a dry run of index_document with document and metadata ##########")
+with open(join(dirname(__file__), '../resources/example.html'), 'r') as document:
+    config = {
+        'retrieve_and_rank': {
+            'dry_run':'true'
+        }
+    }
+    metadata = {
+        'metadata': [
+            {'name':'id', 'value':'12345'}
+        ]
+    }
+    print(json.dumps(document_conversion.index_document(config=config, document=document, metadata=metadata), indent=2))
+
+print("########## Example of a dry run of index_document with document, metadata, and additional config for conversion ##########")
+with open(join(dirname(__file__), '../resources/example.html'), 'r') as document:
+    config = {
+        'convert_document': {
+            'normalized_html': {
+                'exclude_content': {"xpaths":["//body/div"]}
+            }
+        },
+        'retrieve_and_rank': {
+            'dry_run':'true'
+        }
+    }
+    metadata = {
+        'metadata': [
+            {'name':'id', 'value':'12345'}
+        ]
+    }
+    print(json.dumps(document_conversion.index_document(config=config, document=document, metadata=metadata), indent=2))
+
+print("########## Example of index_document with document, metadata (A service instance id, SOLR cluster id, and "
+      "a SOLR collection name must be provided from the Retrieve and Rank service in order to index) ##########")
+with open(join(dirname(__file__), '../resources/example.html'), 'r') as document:
+    config = {
+        'retrieve_and_rank': {
+            'dry_run':'false',
+            'service_instance_id':'YOUR RETRIEVE AND RANK SERVICE INSTANCE ID',
+            'cluster_id':'YOUR RETRIEVE AND RANK SERVICE SOLR CLUSTER ID',
+            'search_collection':'YOUR RETRIEVE AND RANK SERVICE SOLR SEARCH COLLECTION NAME'
+        }
+    }
+    metadata = {
+        'metadata': [
+            {'name':'id', 'value':'12345'}
+        ]
+    }
+    print(json.dumps(document_conversion.index_document(config=config, document=document, metadata=metadata), indent=2))
diff --git a/watson_developer_cloud/document_conversion_v1.py b/watson_developer_cloud/document_conversion_v1.py
index 31430053e..9ad5e29b3 100644
--- a/watson_developer_cloud/document_conversion_v1.py
+++ b/watson_developer_cloud/document_conversion_v1.py
@@ -37,3 +37,16 @@ def convert_document(self, document, config, media_type=None):
         accept_json = config['conversion_target'] == DocumentConversionV1.ANSWER_UNITS
         return self.request(method='POST', url='/v1/convert_document', files=files, params=params,
                             accept_json=accept_json)
+
+    def index_document(self, config, document=None, metadata=None, media_type=None):
+        if document == None and metadata == None:
+            raise AssertionError('Missing required parameters: document or metadata. At least one of those is required.')
+        params = {'version': self.version}
+        files = [('config', ('config.json', json.dumps(config), 'application/json'))]
+        if document != None:
+            filename = os.path.basename(document.name)
+            file_tuple = (filename, document, media_type) if media_type else (filename, document)
+            files.append(('file', file_tuple))
+        if metadata != None:
+            files.append(('metadata', ('metadata.json', json.dumps(metadata), 'application/json')))
+        return self.request(method='POST', url='/v1/index_document', files=files, params=params, accept_json=True)

From f7c791c7de880d4767f70e5ada43f7621b9f6d4a Mon Sep 17 00:00:00 2001
From: "Samir J. Patel" <samirpa@us.ibm.com>
Date: Tue, 5 Jul 2016 13:44:40 -0400
Subject: [PATCH 2/7] [document-conversion] Fixes conditional statement

Fixes a conditional statement to resolve Travis CI failure.
---
 watson_developer_cloud/document_conversion_v1.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/watson_developer_cloud/document_conversion_v1.py b/watson_developer_cloud/document_conversion_v1.py
index 9ad5e29b3..ac1b1bba2 100644
--- a/watson_developer_cloud/document_conversion_v1.py
+++ b/watson_developer_cloud/document_conversion_v1.py
@@ -39,7 +39,7 @@ def convert_document(self, document, config, media_type=None):
                             accept_json=accept_json)
 
     def index_document(self, config, document=None, metadata=None, media_type=None):
-        if document == None and metadata == None:
+        if document is None and metadata is None:
             raise AssertionError('Missing required parameters: document or metadata. At least one of those is required.')
         params = {'version': self.version}
         files = [('config', ('config.json', json.dumps(config), 'application/json'))]

From 0f41152a9f762bb531816b0d836ca453776ddb59 Mon Sep 17 00:00:00 2001
From: "Samir J. Patel" <samirpa@us.ibm.com>
Date: Tue, 5 Jul 2016 16:50:11 -0400
Subject: [PATCH 3/7] [document-conversion] Adds unit tests

This commit adds convert_document and index_document unit tests
for the python sdk project.
---
 resources/simple.html               |  9 +++++
 test/test_document_conversion_v1.py | 54 +++++++++++++++++++++++++++++
 2 files changed, 63 insertions(+)
 create mode 100644 resources/simple.html
 create mode 100644 test/test_document_conversion_v1.py

diff --git a/resources/simple.html b/resources/simple.html
new file mode 100644
index 000000000..1ad0599fc
--- /dev/null
+++ b/resources/simple.html
@@ -0,0 +1,9 @@
+<html>
+<head>
+  <title>Simple HTML Page</title>
+</head>
+<body>
+  <h1>Chapter 1</h1>
+  <p>The content of the first chapter.</p>
+</body>
+</html>
\ No newline at end of file
diff --git a/test/test_document_conversion_v1.py b/test/test_document_conversion_v1.py
new file mode 100644
index 000000000..30475e3f5
--- /dev/null
+++ b/test/test_document_conversion_v1.py
@@ -0,0 +1,54 @@
+# coding=utf-8
+import os
+import responses
+import watson_developer_cloud
+
+
+@responses.activate
+def test_success():
+    convert_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/convert_document'
+    convert_response = '<?xml version="1.0" encoding="UTF-8" standalone="yes"?><html>' \
+                         '<head><title>Simple HTML Page</title></head>' \
+                         '<body><h1>Chapter 1</h1><p>The content of the first chapter.</p></body></html>'
+    document_conversion = watson_developer_cloud.DocumentConversionV1(
+        username="username", password="password")
+
+    responses.add(responses.POST, convert_url,
+                  body=convert_response, status=200,
+                  content_type='application/json')
+
+    with open(os.path.join(os.path.dirname(__file__), '../resources/simple.html'), 'r') as document:
+        config = {'conversion_target': watson_developer_cloud.DocumentConversionV1.NORMALIZED_HTML}
+        document_conversion.convert_document(
+            document=document, config=config, media_type='text/html')
+
+    assert responses.calls[
+               1].request.url == convert_url
+    assert responses.calls[1].response.text == convert_response
+
+    assert len(responses.calls) == 2
+
+    index_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/index_document'
+    index_response = '{"status": "success"}'
+
+    responses.add(responses.POST, index_url,
+                  body=index_response, status=200,
+                  content_type='application/json')
+
+    with open(os.path.join(os.path.dirname(__file__), '../resources/simple.html'), 'r') as document:
+        config = {
+            'retrieve_and_rank': {
+                'dry_run':'false',
+                'service_instance_id':'serviceInstanceId',
+                'cluster_id':'clusterId',
+                'search_collection':'searchCollectionName'
+            }
+        }
+        document_conversion.index_document(
+            config=config, document=document)
+
+    assert responses.calls[
+               1].request.url == index_url
+    assert responses.calls[1].response.text == index_response
+
+    assert len(responses.calls) == 2

From 0de97a50dfb2aaf4984b141b4fd9be76d4fa721f Mon Sep 17 00:00:00 2001
From: "Samir J. Patel" <samirpa@us.ibm.com>
Date: Tue, 5 Jul 2016 17:01:46 -0400
Subject: [PATCH 4/7] [document-conversion] Adds version param

Adds version parameter to the DocumentConversionV1 object
in the unit test.
---
 test/test_document_conversion_v1.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test/test_document_conversion_v1.py b/test/test_document_conversion_v1.py
index 30475e3f5..8f9c6de85 100644
--- a/test/test_document_conversion_v1.py
+++ b/test/test_document_conversion_v1.py
@@ -11,7 +11,7 @@ def test_success():
                          '<head><title>Simple HTML Page</title></head>' \
                          '<body><h1>Chapter 1</h1><p>The content of the first chapter.</p></body></html>'
     document_conversion = watson_developer_cloud.DocumentConversionV1(
-        username="username", password="password")
+        username="username", password="password", version='2015-12-15')
 
     responses.add(responses.POST, convert_url,
                   body=convert_response, status=200,

From c5f74f03ea75793dcc8c93c6519f7c52db85ef3b Mon Sep 17 00:00:00 2001
From: "Samir J. Patel" <samirpa@us.ibm.com>
Date: Wed, 6 Jul 2016 11:59:25 -0400
Subject: [PATCH 5/7] [document-conversion] Fixes unit tests

This commit includes updated fixes for the document
conversion unit tests.
---
 test/test_document_conversion_v1.py | 22 ++++++++--------------
 1 file changed, 8 insertions(+), 14 deletions(-)

diff --git a/test/test_document_conversion_v1.py b/test/test_document_conversion_v1.py
index 8f9c6de85..7cd856e2a 100644
--- a/test/test_document_conversion_v1.py
+++ b/test/test_document_conversion_v1.py
@@ -18,15 +18,11 @@ def test_success():
                   content_type='application/json')
 
     with open(os.path.join(os.path.dirname(__file__), '../resources/simple.html'), 'r') as document:
-        config = {'conversion_target': watson_developer_cloud.DocumentConversionV1.NORMALIZED_HTML}
-        document_conversion.convert_document(
-            document=document, config=config, media_type='text/html')
+        convertConfig = {'conversion_target': watson_developer_cloud.DocumentConversionV1.NORMALIZED_HTML}
+        document_conversion.convert_document(document=document, config=convertConfig, media_type='text/html')
 
-    assert responses.calls[
-               1].request.url == convert_url
-    assert responses.calls[1].response.text == convert_response
-
-    assert len(responses.calls) == 2
+    assert responses.calls[0].request.url == convert_url
+    assert responses.calls[0].response.text == convert_response
 
     index_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/index_document'
     index_response = '{"status": "success"}'
@@ -35,8 +31,8 @@ def test_success():
                   body=index_response, status=200,
                   content_type='application/json')
 
-    with open(os.path.join(os.path.dirname(__file__), '../resources/simple.html'), 'r') as document:
-        config = {
+    with open(os.path.join(os.path.dirname(__file__), '../resources/example.html'), 'r') as document:
+        indexConfig = {
             'retrieve_and_rank': {
                 'dry_run':'false',
                 'service_instance_id':'serviceInstanceId',
@@ -44,11 +40,9 @@ def test_success():
                 'search_collection':'searchCollectionName'
             }
         }
-        document_conversion.index_document(
-            config=config, document=document)
+        document_conversion.index_document(config=indexConfig, document=document)
 
-    assert responses.calls[
-               1].request.url == index_url
+    assert responses.calls[1].request.url == index_url
     assert responses.calls[1].response.text == index_response
 
     assert len(responses.calls) == 2

From 85fa70d7bb70b41a9125401ca9df9c720507fd05 Mon Sep 17 00:00:00 2001
From: "Samir J. Patel" <samirpa@us.ibm.com>
Date: Wed, 6 Jul 2016 12:13:19 -0400
Subject: [PATCH 6/7] [document-conversion] Add version param to url's

Adding the version parameter to the convert and index url's
for the document conversion unit tests.
---
 test/test_document_conversion_v1.py | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/test/test_document_conversion_v1.py b/test/test_document_conversion_v1.py
index 7cd856e2a..fa7bdb6f4 100644
--- a/test/test_document_conversion_v1.py
+++ b/test/test_document_conversion_v1.py
@@ -6,7 +6,7 @@
 
 @responses.activate
 def test_success():
-    convert_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/convert_document'
+    convert_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/convert_document?version=2015-12-15'
     convert_response = '<?xml version="1.0" encoding="UTF-8" standalone="yes"?><html>' \
                          '<head><title>Simple HTML Page</title></head>' \
                          '<body><h1>Chapter 1</h1><p>The content of the first chapter.</p></body></html>'
@@ -24,7 +24,7 @@ def test_success():
     assert responses.calls[0].request.url == convert_url
     assert responses.calls[0].response.text == convert_response
 
-    index_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/index_document'
+    index_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/index_document?version=2015-12-15'
     index_response = '{"status": "success"}'
 
     responses.add(responses.POST, index_url,

From b1660570895e1e63844856a5cefcf54f5e37ef1a Mon Sep 17 00:00:00 2001
From: "Samir J. Patel" <samirpa@us.ibm.com>
Date: Wed, 6 Jul 2016 12:28:58 -0400
Subject: [PATCH 7/7] [document-conversion] Fixes unit tests

Fix for document conversion unit test failure - fixes
assertion of convert & index url's
---
 test/test_document_conversion_v1.py | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/test/test_document_conversion_v1.py b/test/test_document_conversion_v1.py
index fa7bdb6f4..8740cd38e 100644
--- a/test/test_document_conversion_v1.py
+++ b/test/test_document_conversion_v1.py
@@ -6,7 +6,7 @@
 
 @responses.activate
 def test_success():
-    convert_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/convert_document?version=2015-12-15'
+    convert_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/convert_document'
     convert_response = '<?xml version="1.0" encoding="UTF-8" standalone="yes"?><html>' \
                          '<head><title>Simple HTML Page</title></head>' \
                          '<body><h1>Chapter 1</h1><p>The content of the first chapter.</p></body></html>'
@@ -21,10 +21,10 @@ def test_success():
         convertConfig = {'conversion_target': watson_developer_cloud.DocumentConversionV1.NORMALIZED_HTML}
         document_conversion.convert_document(document=document, config=convertConfig, media_type='text/html')
 
-    assert responses.calls[0].request.url == convert_url
+    assert responses.calls[0].request.url == convert_url + '?version=2015-12-15'
     assert responses.calls[0].response.text == convert_response
 
-    index_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/index_document?version=2015-12-15'
+    index_url = 'https://gateway.watsonplatform.net/document-conversion/api/v1/index_document'
     index_response = '{"status": "success"}'
 
     responses.add(responses.POST, index_url,
@@ -42,7 +42,7 @@ def test_success():
         }
         document_conversion.index_document(config=indexConfig, document=document)
 
-    assert responses.calls[1].request.url == index_url
+    assert responses.calls[1].request.url == index_url + '?version=2015-12-15'
     assert responses.calls[1].response.text == index_response
 
     assert len(responses.calls) == 2