Support for JSON

docs/contributing.rst

@@ -75,6 +75,6 @@ Quick start
    <https://github.com/deanmalmgren/flo/blob/master/CONTRIBUTING.md>`__ and send
    pull requests; your help is greatly appreciated!
 
-.. |Build Status| image:: https://travis-ci.org/deanmalmgren/flo.png
-   :target: https://travis-ci.org/deanmalmgren/flo
+.. |Build Status| image:: https://travis-ci.org/deanmalmgren/textract.png
+   :target: https://travis-ci.org/deanmalmgren/textract
 

tests/json/json_is_my_best_friend.json

@@ -0,0 +1,11 @@
+{
+  "hey":"hello world this is some text from a json document", 
+  "should_ignore_this_number": 42, 
+  "should_extract_one_two_three": {
+    "1": "one",
+    "2": "two",
+    "3": "three",
+    "should_be_abcdef": ["a", "b", "c", ["d","e"], {"dict":"f"}]
+  },
+  "finished": "fine"
+}

tests/run_functional_tests.sh

@@ -76,6 +76,7 @@ validate_example ${BASEDIR}/txt/little_bo_peep.txt 1c5fb4478d84c3b3296746e491895
 validate_example ${BASEDIR}/html/snow-fall.html acc2d8c49094e56474006cab3d3768eb
 validate_example ${BASEDIR}/html/what-we-do.html 1fb0263bf62317365cb30246d9e094be
 validate_example ${BASEDIR}/eml/example.eml cb59a5fad8ed8b849e15d53449b1de3f
+validate_example ${BASEDIR}/json/json_is_my_best_friend.json 00a080452bbafa6dfbf81e2a369a4430
 
 # exit with the sum of the status
 exit ${EXIT_CODE}

textract/parsers/__init__.py

@@ -18,8 +18,11 @@ def process(filename, **kwargs):
     # is a relative import so the name of the package is necessary
     root, ext = os.path.splitext(filename)
     ext = ext.lower()
+
+    # cannot call module json.py, conflicts with built-in json library
+    module = '.json_parser' if ext == '.json' else ext
     try:
-        filetype_module = importlib.import_module(ext, 'textract.parsers')
+        filetype_module = importlib.import_module(module, 'textract.parsers')
     except ImportError, e:
         raise exceptions.ExtensionNotSupported(ext)
 

textract/parsers/json_parser.py

@@ -0,0 +1,26 @@
+import json
+
+
+def extract(filename, **kwargs):
+    f = open(filename, 'r')
+    deserialized_json = json.load(f)
+    return get_text(deserialized_json)
+
+
+def get_text(deserialized_json):
+    if isinstance(deserialized_json, dict):
+        result = ''
+        for key in deserialized_json:
+            result += get_text(deserialized_json[key]) + ' '
+        return result
+
+    if isinstance(deserialized_json, list):
+        result = ''
+        for item in deserialized_json:
+            result += get_text(item) + ' '
+        return result
+
+    if isinstance(deserialized_json, basestring):
+        return deserialized_json
+    else:
+        return ''