Cluster the documents by their content, NOT by document ID (eg: person name, location, date).
If you cluster the documents with the name then all the documents containing that name should be near each other as a [login to view URL] can use NLTK toolkit as well or any other algorithms.
You should design interactive features on your visually revealed clusters
[login to view URL] the clusters up and down
2. opening of all document at once
Hello Sir/Ma'am, I have read your project description and i can do this work. Lets have a chat to discuss about its details a bit more. Regards Nouman Ashfaq