Innovate anywhere, anytime Your cloud-based dev studio.

How to Index Binary Files in Django Haystack


Now we are going to index text content which is stored in structured files such as PDFs, Microsoft Office documents, images, etc using haystack and sorl's

In order to read and store the data, we can use SearchBackend.extract_file_contents(self, file_obj) method. It takes the file object, returns a dictionary containing two keys: metadata and contents. The contents value will be a string containing all of the text which the backend managed to extract the file contents.

Here we are overriding NewsIndex prepare method to include the extracted content along with information retrieved from the database:

class NewsIndex(indexes.SearchIndex, indexes.Indexable):
    text = CharField(document=True, use_template=True, analyzer='synonym_analyzer')
    content = indexes.CharField(model_attr='content')

def prepare(self, obj):
        data = super(NewsIndex, self).prepare(obj)
        file_data = self._get_backend(None).extract_file_contents(obj.new_file)
        template = loader.select_template(
            ("search/indexes/proj/new_text.txt", ),
        data["text"] = template.render(Context({
            "object": obj,
            "file_data": var,
        return data

This allows you to insert the extracted text at the appropriate place in your template, 

{{ file_data.contents|striptags|safe }}

To index the documents, we need to generate a schema.xml about our models

./ build_solr_schema >schema.xml

In order to tell sorl about our models schema, just copy the schema.xml and put it in /etc/solr/conf

sudo cp /home/git-projs/elasticproject/schema.xml  /etc/solr/conf

and we can run solr using the following command in which folder you have downloaded 

java -jar start.jar