The Components and how to use them
Dreamcatcher works with four things: Categories, Documents, a Collection, and features.
Categories are used to separate each of the things you write about, these could be categories for subjects and modules you're studying, for each job you're doing for a client, cooking recipes, pretty much anything you want.
Setting Dreamcatcher up for Automatic Filing
The first thing we need to do is create a Collection, this is what will essentially hold your filing system. You could have multiple collections for different purposes (e.g. Study, Work, Creative, Personal), or a single Collection for everything, but whatever you choose, you can easily change your mind later (splitting one collection in to many, merging many collections in to one).
Once we've made a collection, we can either import an existing filing system (selecting the folder containing it e.g. your "Documents" folder), this will create Categories based on all your folders, or we can start manually creating Categories. Here are some I made to organise my lecture notes:
Before you can start automatically filing, Dreamcatcher needs to know a little bit about what belongs in each category. This is done by importing and manually filing a couple of Documents.
Documents can be imported from anywhere on your computer, and Dreamcatcher supports a variety of filetypes (Word Documents, PDF, RTF, TXT and more).
As you choose your documents you can assign them to the categories you made (as illustrated below). If you imported an existing filing system, Dreamcatcher will have already automatically imported all supported documents into their appropriate Categories (folders you had them filed in).
As your Documents are imported, Dreamcatcher reads each document and attempts to automatically retrieve bits of key information (features) and score them with the currently set weighting scheme.
Once you've imported a few documents to some categories, you can take a look at the key information gathered for each document, and how that information was scored by clicking on the document in the Training & Auto-filing window.
The "Training & Auto-filing" tab.
You can also see the information Dreamcatcher deems important/relevant for a particular category. If it seems Dreamcatcher has mistaken some features as important ones, you can easily remove/rectify them. You can also boost the importance of some key terms if you feel it's appropriate.
Now we've done some manual filing, we can test or optimise Dreamcatcher's brain on the information it has. This can be done by clicking "Test" or "Optimise" in the auto filing section of the Training & Auto Filing tab. Dreamcatcher will then give you both a vague and detailed overview on how well it's performing, and give you some quick tips on how best to improve it's performance. It may also perform an automatic optimisation on its self to boost its performance.
To avoid Dreamcatcher putting things in the wrong place, you have the option to set a confidence threshold, whereby any predictions Dreamcatcher makes that fall below your given threshold, won't be filed but instead marked as "Uncategorised", which can later be quickly and easily be manually filed. Once you're happy with everything, you can enable Auto-filing and start introducing documents to Dreamcatcher to be filed automatically.
When you do this, Dreamcatcher will give you a brief summary of how it performed. If you want an explanation as to why, how and where it filed things, and how to improve it's accuracy, just click the "Review Results" button.
Retrieving your Documents
Dreamcatcher keeps hold of everything it's learned about your Documents, which means you can use it as your own personal search engine for your Documents. You can search for things in a variety of interesting ways in the Collection View tab. For example you could just type in to the search bar as you would with Google, or you could right click a document you already have and search for similar documents.
And that's all there is to it!
This application is still a work in progress, and there's lots to do before it's ready for release. Although I intend to release a BETA version soon. If you have any ideas, suggestions, comments, or would be interested in joining the BETA, feel free to get in touch.
The application will be available on:
- Mac OS X