- The server has been redone to remove the ConnNode and now has a single REPLNode.
- The client has been rebuilt (partially) with the gtk3 package.
-
The resource node now connects to a mongo database and uses that database to store user information.
-
Users now have their own REPL nodes.
-
A default user exists in new databases called root, with a default password of entry.
-
Some actions require 'rootAccess'.
-
A :help command lists all existing commands and accompanying information.
- Basic and conditional probability added, the first distribution functions written.
- A simple text-based client has been build, and command line options have been added to Main.
-
ServerNode pulled out of Server.
-
Resource node has been set up, for the moment it only handles logging. Logs are written to ~/.raven/log. Most log event created.
-
:kill now kills the client and sends a kill the command to the sever (not a long term solution, but one that will persist till the client is improved). The client simply stops, but the server sends out kill messages to all nodes, closing them safely.
-
Entry updated. Now requires a summary function (which takes a vector of entries and returns formatted text), a function to return the NA value, and a readEntry function that starts with a normal type. This has not been implemented via read so the Entry can be read and shown as is. Entry also now exports a getEntries function for translating a vector and a countNAs function for the summary system.
-
Table has a function, tableSummary witch uses the summary function exported by entry.
-
BasicEntry and BasicUnboundEntry have been separated to avoid import errors.
-
Stat is a module containing simple statistics functions, all arithmetic (not sampling), with multiple of each as types required.
-
REPL now persists its state via an MVar, all work is done via a node and sent back to the connection (not just an echo sever anymore).
-
Connections are handled by a connection node to smooth communication.
- Echo server setup using Cloud Haskell. Client server interaction done via transportation layer, not via nodes. No internal work yet.
-
Basic REPL set up using the hint library. This is a naive implementation that forces the REPL to be set up again each time.
-
First pass at an Entry class built, with two example data types.
-
Table datatype for data sets build and tested
- Init