diff options
Diffstat (limited to 'megapixels/notebooks/datasets/identity/identity_testing.ipynb')
| -rw-r--r-- | megapixels/notebooks/datasets/identity/identity_testing.ipynb | 50 |
1 files changed, 16 insertions, 34 deletions
diff --git a/megapixels/notebooks/datasets/identity/identity_testing.ipynb b/megapixels/notebooks/datasets/identity/identity_testing.ipynb index 384cca93..3975d0c6 100644 --- a/megapixels/notebooks/datasets/identity/identity_testing.ipynb +++ b/megapixels/notebooks/datasets/identity/identity_testing.ipynb @@ -13,7 +13,7 @@ }, { "cell_type": "code", - "execution_count": 155, + "execution_count": 186, "metadata": {}, "outputs": [], "source": [ @@ -54,23 +54,6 @@ ] }, { - "cell_type": "code", - "execution_count": 159, - "metadata": {}, - "outputs": [ - { - "name": "stdout", - "output_type": "stream", - "text": [ - "/media/adam/ah8tb/work/megapixels_dev/env/google_knowledge_graph_api.env\n" - ] - } - ], - "source": [ - "print(app_cfg.FP_KNOWLEDGE_GRAPH_ENV)" - ] - }, - { "cell_type": "markdown", "metadata": {}, "source": [ @@ -79,7 +62,7 @@ }, { "cell_type": "code", - "execution_count": 160, + "execution_count": 188, "metadata": {}, "outputs": [ { @@ -92,7 +75,8 @@ } ], "source": [ - "names = identity_utils.get_names(types.Dataset.LFW)\n", + "names = identity_utils.get_names(types.Dataset.\n", + " )\n", "print(names['names_query'][0:10])\n", "print(names['names_orig'][0:10])" ] @@ -108,14 +92,12 @@ }, { "cell_type": "code", - "execution_count": 161, + "execution_count": 164, "metadata": {}, "outputs": [], "source": [ "# read API key\n", - "\n", - "api_key = open(app_cfg.FP_KNOWLEDGE_GRAPH_ENV).read()\n", - "kg_api = api_utils.GoogleKnowledgeGraph(api_key)\n", + "kg_api = api_utils.GoogleKnowledgeGraph()\n", "wp_api = api_utils.WikipediaAPI()" ] }, @@ -128,25 +110,23 @@ }, { "cell_type": "code", - "execution_count": 128, + "execution_count": 165, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "wp\n", + "wp----\n", "https://en.wikipedia.org/w/api.php?redirects=&ppprop=displaytitle&prop=pageprops%7Cpageimages%7Cdescription&generator=prefixsearch&action=query&format=json&piprop=thumbnail&pilimit=1&gpssearch=Vicente+Fox&gpsnamespace=0&gpslimit=1\n", "{'wp_accessed': True,\n", " 'wp_description': 'President of Mexico',\n", " 'wp_name': 'Vicente Fox',\n", " 'wp_page_id': '32836'}\n", - "kg\n", + "kg----\n", "{'kg_accessed': True,\n", - " 'kg_bio': 'Vicente Fox Quesada, RSerafO is a Mexican businessman and '\n", - " 'politician who served as the 55th President of Mexico from 1 '\n", - " 'December 2000 to 30 November 2006.\\n',\n", - " 'kg_bio_url': 'https://en.wikipedia.org/wiki/Vicente_Fox',\n", + " 'kg_bio': '',\n", + " 'kg_bio_url': '',\n", " 'kg_description': 'Former President of Mexico',\n", " 'kg_error': '',\n", " 'kg_id': '/m/081f4',\n", @@ -174,14 +154,15 @@ }, { "cell_type": "code", - "execution_count": 162, + "execution_count": 168, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ - "0.7714285714285716\n" + "0.7714285714285716\n", + "0.7142857142857143\n" ] } ], @@ -189,7 +170,8 @@ "#print(identity_utils.names_match('Andréss Iniestas', 'Andres Iniestalossas Jr.', as_float=True))\n", "#print(identity_utils.names_match('Adoor Gopalakrishnan', 'Adoors Gopalakarishnan', as_float=True))\n", "#print(identity_utils.names_match('Dave Letterman', 'David Letterman', as_float=True))\n", - "print(identity_utils.names_match('Charles Dickens', 'Charles Boooker', as_float=True))\n", + "print(identity_utils.names_match('Charles Dickens', 'Charles Boooker', as_float=True, compound_score=True))\n", + "print(identity_utils.names_match('Charles Dickens', 'Charles Boooker', as_float=True, compound_score=False))\n", "#print(identity_utils.names_match('Donald Trump', 'Donald J. Trump', as_float=True))\n", "#print(identity_utils.names_match('Wang Fei', 'Fei Wang III', as_float=True))" ] |