Release alpha 0.3

Code:
* Rewritten contributions by previous team member.
* Added icon file chooser

Documentation:
* Added book version with some academic requirements
* Date and version automatically get inputted on document
* Added implementation section
* Expanded previous chapters

.gitignore

@@ -1,6 +1,9 @@
 # PDF files
 *.pdf
 
+# Word files
+*.docx
+
 # TeX files
 *.tex
 

.travis.yml

@@ -7,6 +7,7 @@ install: true
 
 script: true
 
+
 after_success:
     - sudo apt-get install texlive-latex-extra
     - cd docs
@@ -18,6 +19,13 @@ after_success:
     - cd .. && mv standalone/* .
     - make travis && cd ..
 
+# Run tests on master & dev
+branches:
+    only:
+        - master
+        - dev
+
+# Only deploy on master
 deploy:
     provider: releases
     api_key: "$GH_TOKEN"
@@ -28,9 +36,6 @@ deploy:
         branch:
             - master
 
-# Only run CI on master
-branches:
-    only:
-        - master
-        - dev
-
+# Stop bothering me
+notifications:
+    email: false

docs/Makefile

@@ -1,48 +1,69 @@
 PDF := persimmon.pdf  # PDF Main Target
+MARKDOWN := introduction.md state_of_the_art.md objectives.md risk_analysis.md \
+	implementation.md postmortem.md  # Markdown files
+BODY := body.tex  # Markdown files will be converted to this intermediate step
+#APPENDICES := appendixX.md  # Markdown Appendices
+#APPENDIX := appendix.tex  # And appendices to this intermediate step
+# METADATA := metadata.yaml  # Metadata files (Author, Date, Title, etc..)
+BIBLIOGRAPHY := persimmon.bib  # BibLaTeX bibliography
 CSL := emerald-harvard.csl  # CSL file used for citations
 TEMPLATE := template.tex  # LaTeX template for producing PDF
-BIBLIOGRAPHY := persimmon.bib  # BibLaTeX bibliography
-MARKDOWN := chapter0.md chapter1.md chapter2.md chapter3.md  # Markdown files
-GRAPHS := $(wildcard graphs/*.tex)
-IMAGES := $(GRAPHS:.tex=.pdf)
-IMAGES += $(wildcard graphs/*.png)  # Standalone pictures to be inserted
-LATEXHEADERS := latexheaders.tex  # Additional LaTeX headers
-METADATA := metadata.yaml  # Metadata files (Author, Date, Title, etc..)
-
-all: pdf
-
-pdf: $(MARKDOWN) $(BIBLIOGRAPHY) $(CSL) $(TEMPLATE) $(IMAGES) $(METADATA)
-	pandoc --standalone --smart --latex-engine xelatex --template $(TEMPLATE) \
-		--top-level-division chapter --bibliography $(BIBLIOGRAPHY) --csl $(CSL) \
-		--include-in-header $(LATEXHEADERS) $(METADATA) $(MARKDOWN) -o $(PDF)
-
-travis: $(IMAGES)
-	pandoc --standalone --smart --latex-engine xelatex --template $(TEMPLATE) \
-		--chapters --bibliography $(BIBLIOGRAPHY) --csl $(CSL) \
-		--include-in-header $(LATEXHEADERS) $(METADATA) $(MARKDOWN) -o $(PDF)
-
-# For standalone images (Not used)
+
+GRAPHS := $(wildcard graphs/*.tex)  # Latex diagrams
+IMAGES := $(wildcard graphs/*.png)  # .png images
+IMAGES += $(GRAPHS:.tex=.pdf)  # Generated PDF Images
+
+all: $(PDF)
+
+# Main PDF, travis ci and book to print version
+$(PDF): $(BODY) $(TEMPLATE) $(IMAGES) # TODO: Add abstract
+	pandoc --smart --standalone --latex-engine xelatex --template $(TEMPLATE) \
+		--metadata author:"Álvaro Bermejo" \
+		--metadata date:"$(shell date +"%d/%m/%Y") ($(shell git describe --abbrev=0 --tags))" \
+		--metadata title:"Persimmon" --metadata fontsize:"12pt" --toc \
+		--metadata subtitle:"A scikitlearn visual programming interface" \
+		--metadata mainlang:"English" --metada keywords:"Machine Learning","Visual Programming" \
+		--metadata papersize:"A4" --metadata sansfont:"Helvetica Neue LT Com" \
+		--metadata colorlinks --metadata documentclass:"scrreprt" \
+		--top-level-division chapter $(BODY) -o $@
+
+# Main PDF, travis ci and book to print version
+travis: $(BODY) $(APPENDIX) $(TEMPLATE) $(IMAGES)
+	pandoc --smart --standalone --latex-engine xelatex --template $(TEMPLATE) \
+		--metadata author:"Álvaro Bermejo" \
+		--metadata date:"$(shell date +"%d/%m/%Y") ($(shell git describe --abbrev=0 --tags))" \
+		--metadata title:"Persimmon" --metadata fontsize:"12pt" --toc \
+		--metadata subtitle:"A sklearn visual programming interface" \
+		--metadata mainlang:"English" --metada keywords:"Machine Learning","Visual Programming" \
+		--metadata papersize:"A4" \
+		--metadata colorlinks --metadata documentclass:"scrreprt" \
+		--chapters $(BODY) $(APPENDIX) -o $(PDF)
+
+
+book_complu: $(BODY) $(APPENDIX) $(TEMPLATE) $(IMAGES)
+	pandoc --smart --standalone --latex-engine xelatex --template $(TEMPLATE) \
+		--metadata author:"Álvaro Bermejo" --metadata date:"Director: Pablo Moreno Ger" \
+		--metadata title:"Persimmon" --metadata fontsize:"12pt" --toc \
+		--metadata subtitle:"A scikitlearn visual programming interface" \
+		--metadata mainlang:"English" \
+		--metadata papersize:"A4" --metadata sansfont:"Helvetica Neue LT Com" \
+		--metadata documentclass:"scrbook" --metadata institute:"Universidad Complutense" \
+		--top-level-division chapter $(BODY) $(APPENDIX) -o bool_$(PDF)
+
+
+# For standalone images 
 graphs/%.pdf: graphs/%.tex
-	xelatex $<
-
-## Splitted creation (Not currently working)
-#CHAPTERS := $(MARKDOWN:.md=.tex)  # LaTeX Chapters
-#GRAPHS := $(wildcard graphs/*.tex)
-#IMAGES += $(GRAPHS:.tex=.pdf)
-#
-##splitted: $(CHAPTERS) $(BIBLIOGRAPHY) $(CSL) $(TEMPLATE)
-#	pandoc --standalone --smart --latex-engine xelatex --template $(TEMPLATE) \
-#		--top-level-division chapter --bibliography $(BIBLIOGRAPHY) --csl $(CSL) \
-#	   	$(CHAPTERS) -o $(PDF)
-#
-#%.tex: metadata.yaml %.md
-#	pandoc --no-tex-ligatures metadata.yaml $*.md -o $@
-#
-## For standalone images (Not used)
-#graphs/%.pdf: graphs/%.tex
-#	xelatex $<
-#
+	xelatex $< > /dev/null # TODO: actually output in graphs directory
+	mv $*.pdf graphs/
+
+# Body and Appendices Middle Steps creation
+$(BODY): $(MARKDOWN)
+	pandoc --no-tex-ligatures --bibliography $(BIBLIOGRAPHY) --csl $(CSL) \
+		metadata.yaml $(MARKDOWN) -o $@
+
+$(APPENDIX): $(APPENDICES)
+	pandoc --no-tex-ligatures $(APPENDICES) -o $@
 
 clean:
-	rm -f *.pdf chapter?.tex *.log *.aux *.png
+	rm -f $(BODY) $(APPENDIX) graphs/*.pdf *.pdf *.log *.aux
 

docs/appendixX.md

@@ -0,0 +1,30 @@
+Appendix X: How was this document made?
+=======================================
+
+This document was written on Markdown, and converted to PDF
+using Pandoc.
+
+Process
+-------
+Document is written on Pandoc's extended Markdown, and can be broken amongst
+different files. Images are inserted with regular Markdown syntax for images.
+A YAML file with metadata information is passed to pandoc, containing things 
+such as Author, Title, font, etc... The use of this information depends on
+what output we are creating and the template/reference we are using.
+
+
+Diagrams
+--------
+Diagrams are were created with LaTeX packages such as tikz or pgfgantt, they 
+can be inserted directly as PDF, but if we desire to output to formats other
+than LaTeX is more convenient to convert them to .png filesi with tools such
+as `pdftoppm`.
+
+
+References
+------------
+References are handled by pandoc-citeproc, we can write our bibliography in
+a myriad of different formats: bibTeX, bibLaTeX, JSON, YAML, etc..., then
+we reference in our markdown, and that reference works for multiple formats
+
+

docs/graphs/objectives.tex

@@ -9,7 +9,7 @@
 		\draw (-1, 2.8) rectangle (1, 3.5) node[midway, gray] {Parity};
 		\draw [->, gray, thick] (0, 3.6) -- (0, 4.3);
 		\draw (-1, 4.4) rectangle (1, 5.1) node[midway, gray] {Compilation};
-		\draw [red, dashed] (-3, 5.5) -- (3, 5.5) node[below left] {Out of scope};
+		\draw [red, dashed] (4, 5.35) -- (-4, 5.35) node[above right] {Out of scope};
 		\draw [->, gray, thick] (-0.1, 5.2) -- (-2, 5.9);
 		\draw [->, gray, thick] (0.1, 5.2) -- (2, 5.9);
 		\draw (-3, 6) rectangle (-1, 6.7) node[midway, gray] {Web};

docs/implementation.md

@@ -0,0 +1,11 @@
+Implementation
+==============
+
+First Iteration
+---------------
+
+![Early "static" interface](graphs/early_interface.png)
+
+![File chooser](graphs/filechooser.png)
+
+

docs/introduction.md

@@ -0,0 +1,48 @@
+Introduction
+============
+
+Description
+-----------
+Persimmon is a visual programming interface for sklearn.
+
+This projects involves a variety of Computer Science topics, such as User 
+Experience (Main topic as the project is driven by the users feedback and 
+engagement with the project), Machine Learning (We don't write the algorithms,
+but need extensive knowledge of them to surface all their options) Software 
+Engineering (We have to interact with already built software, using interfaces 
+and organizing code through object-oriented techniques), Compilers (Language 
+parsing and transpilers) and a number of tangentially related topics such as 
+Machine Learning, I/O, preprocessing of data, etc.
+
+Motivation
+----------
+After learning about Machine Learning on university last year I was able to get 
+an internship working for a company on the algorithmic trading sector..
+
+There, amongst other duties, I aided with moving the codebase from MATLAB to 
+Python, and during that process I realised many of my co-workers struggled with
+the switch. All of the were not computer scientists, but instead came from a 
+variety of backgrounds such as Maths, Physics, Electric Engineering, 
+Statistics or Aerospacial Engineering.
+
+Yet they were the whole of the department, their work requires a very high 
+level of theoretical maths knowledge, and so happens that these experts from
+these fields tend to not have a lot of general programming skills, they mostly 
+work with specialized languages, tailored to these tasks such as MATLAB, R, 
+Julia, etc, and moving to a general purpose language such as Python involves 
+learning about a plethora of additional topics, such as Object Oriented 
+Programming, custom complex Datastructures or CPU cache optimization.
+
+The situation is even more complicated for newcomers to Machine Learning, as 
+they not only have the programming barrier but also have to overcome the 
+difficulties of the algorithms themselves, something Computer Scientists also 
+struggle with (In many cases even more because their weaker maths skills).
+
+So this project serves a double purpose, it helps with the programming barrier,
+and it aids with the Machine Learning process as it allows the learner to focus
+on the connections, intuitions and mathematical basis and not on the 
+implementation details and the quirks of the concrete language.
+
+This hypothesis that visual learning can improve understanding is supported by 
+numerous sources such as [@fry2007visualizing] and [@principles].
+

docs/metadata.yaml

@@ -1,10 +1,18 @@
 
 ---
-author: "Álvaro Bermejo"
+author: Álvaro Bermejo
+date: 2017-01-01
 title: "Persimmon: a visual interface for sklearn"
-papersize: "A4"
-fontsize: "12pt"
-mainlang: "English"
-toc: yes
+subtitle: A scikit-learn visual prgramming interface
+version: 0.2
+sansfont: Helvetica Neue LT Com
+colorlinks: True
+documentclass: scrreprt
+institute: Universidad Complutense and University of Hertforshire
+papersize: A4
+fontsize: 12pt
+mainlang: English
+toc: True
+top-level-division: chapter
 ---
 

docs/chapter2.md → docs/objectives.md

@@ -1,11 +1,12 @@
 Objectives
 ==========
-The best way we can describe the project is by dividing the objectives.
-and the best way to understand the progression of those and their relation
+
+The best way to describe the project is by dividing the objectives.
+And the best way to understand the progression of those and their relation
 is with a diagram.
 
 
-![Objectives Tree](objectives.pdf)
+![Objectives Tree](graphs/objectives.pdf)
 
 **Capped** is more than a minimum viable product, a extensive proof-of-concept, 
 with a few limited algorithms and the ability of inputing `.csv` files. with a
@@ -16,21 +17,26 @@ buttons.
 interaction. we don't really care much about having the same number of 
 underlying algorithms because that's not the focus of the project.
 
-And the final objective is **compilation**, the ability to get the python 
+And the final objective is **Compilation**, the ability to get the python 
 source code from the visual representation. also improving the interface to
 have a better flow, such as in unreal blueprints, which provide a very 
 intuitive interface [@shah2014mastering].
 
+This milestone would bring Persimmon utility outside just the realm of 
+learning, as it would be a convenince tool for the exploratory work of any 
+ML solution (Business case, a Kaggle[^kaggle] competition, etc...
+
 Out of scope, but further applications of the system are **web/junyper**
 integration that means the system would be accesible from a website interface,
 and script **synthesization**, which is the opposite of compilation, meaning
 the ability to visualize on persimmon a python source file.
 
 Now that we understand the objectives we can draw a much detailed gantt diagram.
 
-![Gantt Diagram](gantt.pdf)
+![Gantt Diagram](graphs/gantt.pdf)
 
 
 We ommited previous months that included idea refinement but are not 
 interesting for us.
 
+[^kaggle]: [Kaggle.com](https://www.kaggle.com/)

docs/postmortem.md

@@ -0,0 +1,6 @@
+Postmortem
+==========
+
+Bibliography
+============
+

docs/chapter3.md → docs/risk_analysis.md

@@ -22,7 +22,7 @@ Prevention & Mitigation
 -----------------------
 
 | Risk Factor  |     Low Impact     |     Medium Impact    |    High Impact   |
-|--------------|--------------------|----------------------|------------------|
+|:------------ |:------------------:|:--------------------:|:----------------:|
 | Requirements | Not defined enough | Change at late stage | Unreachable goal |
 |  Technology  | Performance issues |   Interoperability   |   Major errors   |
 
@@ -52,6 +52,3 @@ analysis of the capabilities of the platform must be done before starting the
 project, identifying possible faults and providing possible solutions and or 
 alternatives.
 
-
-Bibliography
-============