Sharing models and code made easy

New Python and Octave code can be submitted to the repository directly from TinkerCell. The "Save" button in TinkerCell's coding window will do the work for the user. Basically, it will generate the SVN command for uploading to the repository (the user must have subversion installed). The "tinkercelluser" sourceforge account is used for the upload process.




It is also simple for a user to submit models to the repository:

File -> Export -> Share model


The above dialog appears, asking the user for the name and family of the model. The interface at present does not allow the user to create new families of models; that will require modifying the modules.xml file and populating (correctly) the participant types and roles, which might be a bit too much for most users. So adding new families is kept as something that only more advanced users would do (and they would directly edit the xml file)

Graphical tooltips for sub-models and functions

The Module Tool now shows a little diagram at the bottom right whenever the user hovers above a connection that has a hidden model inside. The image is regenerated whenever the inner model changes.

The decorator icon is also modified so that it is consistently the same icon for all connections that have a hidden model inside. This is due to confusions caused by the inconsistent meanings of the other decorators.

In the next screenshot, the tool-tip window displays the function that is used to assign the value to the repressor binding site. This is important because TinkerCell allows different modeling strategies, so the user (quickly) should know when an assignment rule is used.

TinkerCell symbols

Doc with all the TinkerCell symbols:

https://docs.google.com/document/d/1O2JLVdBRDheurURNbPCtmGwwWC6h_JoS-hnWrA208x4/edit?hl=en

PDF export finally works

Finally resolved the long postponed issue of PDF export. The solution involved changing the ConnectionGraphicsItem class from being a QGraphicsPathItem to a QGraphicsItemGroup. The group then contains path items with it. There are three path items -- one for the main connection itself, one for a white border around it (when lines overlap), and one for the red border when connections are selected. Why does this work? No idea. There is something about the way Qt renders graphics path items that I am unaware of, so its probably best to use their available classes (as in this new strategy).

The exported PDF (from "screenshot" button) is clearly using vector graphics and not pixels.

Small change to assignment rules

Small change to the way assignment rules are declared. If the "assignments" table contains a row called "self", then that row is assumed to be the assignment rule for that component. Before the row name was the component's name, but that was causing problems when sub-processes were merged with the host model, i.e. renaming issues.

And parts on plasmid look better now (side bars become invisible)

Automatically create and upload installer

In Windows and Mac, it is a bit time consuming to create the TinkerCell installer and upload it to Sourceforge.

As a solution, I created a simple CMake-generated script that does the work. Different scripts are generated based on the operating system. The Windows and Linux scripts build the program and upload it to sourceforge. Linux script can also upload the source code. The Mac script does not upload yet, but it does generate the image file (which is most of the work). For developers who want to build TinkerCell on their own machines, new instructions are available in the README.txt file included with the code. Here is a summary:

1. (windows) Install Inno setup, ISTools, and WinSCP.. and CMake of course

2. Download TinkerCell source code

3. Run CMake GUI (use the "grouped" view for convenience)

4. Set the TinkerCell source folder (/trunk) and the build folder (/trunk/BUILD)

5. Select TinkerCell lite versions if desired (under the TinkerCell group)

6. Run "Configure" and "Generate". Look for the message in red, which will tell you where the installation script is created

7. Run the installation script. It will create the program, the installer, and upload it!

Steps 1-6 only need to be run the first time when setting up CMake

For Mac...all the steps are the same, except for #7

7. Run the create_bundled_app.sh script in the BUILD/bin folder

For Linux, there is the additional option of selecting 64-bit or 32-bit OS. The is also the option of uploading the source along with the binaries...

7. Run upload.sh

TinkerCell "lite" revisited

Introducing TinkerCellScript, which is currently using Antimony scripts (but the scripting framework is open to other parsers...)




... and TinkerCellLite, which is just for drawing DNA-based designs

Integrated COPASI and sliders

The COPASI C++ API has been integrated into TinkerCell. Simulation functions, MCA functions, and steady state analysis function have been added. The image above shows a plot of the Eigenvalues with sliders to visualize how the Eigenvalues change due to change in parameters.

A COPASI C API has been created. The API uses Hash tables for quick look-up, thus adding the convenience of name-based look-up. See copasi_api.h for the functions.

Circular Plasmids

The above picture was the old version (not good). Below is the "new" version (much better because the black bars are removed). The "upstream" and "downstream" calculations are performed by taking the starting node and walking along the circle in increments of 0.02 * PI degrees (i.e. 100 points on the circle). If any part is smaller than that (hope not), it might get skipped. The AutoGeneRegulatoryTool now has two functions -- allPartsLinear and allPartsCircular -- one is used for linear arrangement of parts and the other for circular.

Octave Interpreter issues in Windows

On Windows, there is an issue when passing Swig types into Octave functions. Since Octave works with matrices, converting tc_matrix to Octave matrix is most important. The "fromTC" function does the job in Linux, but in windows, the OctaveInterpreter does a regex-replace of fromTC and replaces it with a double-for-loop that converts the tc_matrix into Octave's matrix.

Also, setInterpreter method has been added to the ConsoleWindow class. Here is how to use it (taken from SimpleDesigner.cpp):

ConsoleWindow * console = mainWindow.console();
OctaveInterpreterThread * octaveInterpreter = new OctaveInterpreterThread("tinkercell.oct", "libtcoct", &mainWindow);
octaveInterpreter->initialize();
console->setInterpreter(octaveInterpreter);