Merge pull request #312 from JamesPHoughton/split-parsing-building

PySD v3.0 (Split parsing and building)

.gitignore

@@ -11,3 +11,4 @@ tests/cover/
 tests/htmlcov/
 .idea/*
 docs/_build/*
+docs/tables/*.csv

MANIFEST.in

@@ -1,4 +1,4 @@
 include requirements.txt
 include README.md
 include LICENSE
-include pysd/translation/xmile/smile.grammar
+graft pysd/translators/*/parsing_grammars

README.md

@@ -1,5 +1,6 @@
 PySD
 ====
+
 [![Coverage Status](https://coveralls.io/repos/github/JamesPHoughton/pysd/badge.svg?branch=master)](https://coveralls.io/github/JamesPHoughton/pysd?branch=master)
 [![Anaconda-Server Badge](https://anaconda.org/conda-forge/pysd/badges/version.svg)](https://anaconda.org/conda-forge/pysd)
 [![PyPI version](https://badge.fury.io/py/pysd.svg)](https://badge.fury.io/py/pysd)
@@ -13,16 +14,14 @@ Simulating System Dynamics Models in Python
 
 This project is a simple library for running [System Dynamics](http://en.wikipedia.org/wiki/System_dynamics) models in python, with the purpose of improving integration of *Big Data* and *Machine Learning* into the SD workflow.
 
-**The current version needs to run at least Python 3.7. If you need support for Python 2, please use the release here: https://github.com/JamesPHoughton/pysd/releases/tag/LastPy2**
-
-**table2py feature was dropped in version 2.0.0, please use the release here if you want to build PySD model from a tabular file: https://github.com/JamesPHoughton/pysd/releases/tag/v1.11.0**
+**The current version needs to run at least Python 3.7.**
 
 ### Resources
+
 See the [project documentation](http://pysd.readthedocs.org/) for information about:
 
 - [Installation](http://pysd.readthedocs.org/en/latest/installation.html)
-- [Basic Usage](http://pysd.readthedocs.org/en/latest/basic_usage.html)
-- [Function Reference](http://pysd.readthedocs.org/en/latest/functions.html)
+- [Getting Started](http://pysd.readthedocs.org/en/latest/getting_started.html)
 
 For standard methods for data analysis with SD models, see the  [PySD Cookbook](https://github.com/JamesPHoughton/PySD-Cookbook), containing (for example):
 
@@ -38,7 +37,6 @@ You can also cite the library using the [DOI provided by Zenodo](https://zenodo.
 
 [![DOI](https://zenodo.org/badge/DOI/10.5281/zenodo.5654824.svg)](https://doi.org/10.5281/zenodo.5654824)
 
-
 ### Why create a new SD simulation engine?
 
 There are a number of great SD programs out there ([Vensim](http://vensim.com/), [iThink](http://www.iseesystems.com/Softwares/Business/ithinkSoftware.aspx), [AnyLogic](http://www.anylogic.com/system-dynamics), [Insight Maker](http://insightmaker.com/), and [others](http://en.wikipedia.org/wiki/List_of_system_dynamics_software)). In order not to waste our effort, or fall victim to the [Not-Invented-Here](http://en.wikipedia.org/wiki/Not_invented_here) fallacy, we should have a very good reason for starting a new project.

docs/Makefile

@@ -48,9 +48,12 @@ help:
 	@echo "  doctest    to run all doctests embedded in the documentation (if enabled)"
 	@echo "  coverage   to run coverage check of the documentation (if enabled)"
 
-clean:
+clean: clean_tables
 	rm -rf $(BUILDDIR)/*
 
+clean_tables:
+	rm -f tables/*.csv
+
 html:
 	$(SPHINXBUILD) -b html $(ALLSPHINXOPTS) $(BUILDDIR)/html
 	@echo

docs/development/about.rst → docs/about.rst

@@ -8,25 +8,25 @@ The last few years have witnessed a massive growth in the collection of social a
 
 So far, however, these new techniques are largely confined to variants of statistical summary, categorization, and inference; and if causal models are used, they are generally static in nature, ignoring the dynamic complexity and feedback structures of the systems in question. As the field of data science matures, there will be increasing demand for insights beyond those available through analysis unstructured by causal understanding. At that point data scientists may seek to add dynamic models of system structure to their toolbox.
 
-The field of system dynamics has always been interested in learning about social systems, and specializes in understanding dynamic complexity. There is likewise a long tradition of incorporating various forms of data into system dynamics models.3 While system dynamics practice has much to gain from the emergence of new volumes of social data, the community has yet to benefit fully from the data science revolution.
+The field of system dynamics has always been interested in learning about social systems, and specializes in understanding dynamic complexity. There is likewise a long tradition of incorporating various forms of data into system dynamics models. While system dynamics practice has much to gain from the emergence of new volumes of social data, the community has yet to benefit fully from the data science revolution.
 
 There are a variety of reasons for this, the largest likely being that the two communities have yet to commingle to a great extent. A further, and ultimately more tractable reason is that the tools of system dynamics and the tools of data analytics are not tightly integrated, making joint method analysis unwieldy. There is a rich problem space that depends upon the ability of these fields to support one another, and so there is a need for tools that help the two methodologies work together. PySD is designed to meet this need.
 
 
 General approaches for integrating system dynamic models and data analytics
 ---------------------------------------------------------------------------
-Before considering how system dynamics techniques can be used in data science applications, we should consider the variety of ways in which the system dynamics community has traditionally dealt with integration of data and models. 
+Before considering how system dynamics techniques can be used in data science applications, we should consider the variety of ways in which the system dynamics community has traditionally dealt with integration of data and models.
 
 The first paradigm for using numerical data in support of modeling efforts is to import data into system dynamics modeling software. Algorithms for comparing models with data are built into the tool itself, and are usable through a graphical front-end interface as with model fitting in Vensim, or through a programming environment unique to the tool. When new techniques such as Markov chain Monte Carlo analysis become relevant to the system dynamics community, they are often brought into the SD tool.
- 
+
 This approach appropriately caters to system dynamics modelers who want to take advantage of well-established data science techniques without needing to learn a programming language, and extends the functionality of system dynamics to the basics of integrated model analysis.
 
-A second category of tools uses a standard system dynamics tool as a computation engine for analysis performed in a coding environment. This is the approach taken by the Exploratory Modeling Analysis (EMA) Workbench6, or the Behavior Analysis and Testing Software (BATS)7. This first step towards bringing system dynamics to a more inclusive analysis environment enables many new types of model understanding, but imposes limits on the depth of interaction with models and the ability to scale simulation to support large analysis.
+A second category of tools uses a standard system dynamics tool as a computation engine for analysis performed in a coding environment. This is the approach taken by the Exploratory Modeling Analysis (EMA) Workbench, or the Behavior Analysis and Testing Software (BATS). This first step towards bringing system dynamics to a more inclusive analysis environment enables many new types of model understanding, but imposes limits on the depth of interaction with models and the ability to scale simulation to support large analysis.
+
+A third category of tools imports the models created by traditional tools to perform analyses independently of the original modeling tool. An example of this is SDM-Doc, a model documentation tool, or Abdel-Gawad et. al.’s eigenvector analysis tool. It is this third category to which PySD belongs.
 
-A third category of tools imports the models created by traditional tools to perform analyses independently of the original modeling tool. An example of this is SDM-Doc8, a model documentation tool, or Abdel-Gawad et. al.’s eigenvector analysis tool9. It is this third category to which PySD belongs. 
-
 The central paradigm of PySD is that it is more efficient to bring the mature capabilities of system dynamics into an environment in use for active development in data science, than to attempt to bring each new development in inference and machine learning into the system dynamics enclave.
 
-PySD reads a model file – the product of a modeling program such as Vensim10 or Stella/iThink11 – and cross compiles it into Python, providing a simulation engine that can run these models natively in the Python environment. It is not a substitute for these tools, and cannot be used to replace a visual model construction environment.
+PySD reads a model file – the product of a modeling program such as Vensim or Stella/iThink – and cross compiles it into Python, providing a simulation engine that can run these models natively in the Python environment. It is not a substitute for these tools, and cannot be used to replace a visual model construction environment.