Use python matplotlib to create 3D vector-based graphics, 3D vector graphics DIY

I drew the following graphs using Python and it's Matplotlib. If you are interested in the python code, they are here (not including the first gyroscope figure, will be added later). I am trying to improve the code into many functions/modules so that drawing can be convenient and automated. The part related to plotting starts from line 128, because this is just a portion of another bigger project. Just want to show this is not as time-consuming as one may think it is. We can create or draw beautiful vector-based graphs without buying expensive software like Adobe Illustrator or Acdsee Canvas. However as you can guess with this method you will need the numerical x,y,z locations of everything! (which is generated from yet another portion of the program!)

Aug 2015 added

The python code to generate this fig is here.

----

2016/Jan 15

教illustrator不如教3d computer drawing

簡單2D畫圖->小畫家(點陣圖)缺點:放大會糊掉

向量2D繪圖Inkscape(向量圖)放大不會糊掉

但大家都知道，若要畫立體的東西，在平面上作圖是比較困難的，你的3d空間感要很好。

複雜的立體物件就更難在平面上畫出來了，這時候就要其他有效率的方式。

但常常我們要畫的東西都是立體的阿

電腦輔助繪圖computer aided drawing/design

-> 3d繪圖空間->讓你可以在3d空間中作畫

可以在電腦建立的3D虛擬空間中作圖

一般這種電腦輔助繪圖軟體

AutoCAD, Maya, 3ds Max 等等

很貴!!

有沒有免費的方法呢?有很多

也歡迎大家舉薦

我這邊野人獻曝的跟大家分享我使用的免費方法

用python程式語言作3d繪圖

利用python的3d繪圖指令庫matplotlib

可以畫出讓你滿意的3d向量圖

此方法並非傳統的所見及所得

而是所見及所要

比如說

你要一個線段 是要給個指令像這樣draw_line(pointA,pointB)

然後執行後才會看到結果

所以得要知道pointA,pointB的位置

那AutoCAD的3D繪圖軟體Inventor的話就是所見及所得

先在3D空間中訂出你要的繪圖平面(是2D的)然後再到這個繪圖平面上繪圖。

所以基本上還是在2D平面上繪圖，只是你可以變換你的繪圖平面

你可以想像的到這方法會有一些限制

因為每次你都必須在一個2D平面上作圖

不過好處是畫出來你可以馬上看到

一篇介紹各種各樣向量繪圖軟體的文章，整理得相當好。

https://www.zhihu.com/question/21664179

3D動畫師(地域級)

http://www.bugman123.com/FluidMotion/index.html

待續...

-----

2014/July

由於不想買好貴好貴的向量繪圖軟體，也不想用盜版的，因為現在有些學校提供的防毒軟體會去殺盜版軟體的安裝檔，因此現在要用盜版的也不容易啊。用了一下免費的Inkscape發現並不適合我需求，我需求是必須能有在三度空間中繪圖的功能，見以上第二圖。免費的3D向量編輯軟體好像是有一些，但與其花那個時間去學，我決定還是土法煉鋼自己來試試看吧(也想要能夠擺脫連小地方都一直在倚靠仰賴歐美的好用軟體產品，我們自己也應該要能在小地方慢慢獨立。)因此我用python程式語言來畫3D三維的向量圖形，最終在輸出成平面圖檔如png，平面向量圖檔pgf,eps等。發現其實並沒那麼花時間。把一些常用的功能，畫線段，畫弧線的指令，寫出來，並且只要朝向指令重複使用化的方向去寫code，基本上只要寫一次，以後就剪下貼上嚕。這邊把code分享給大家，希望大家試試看，或提供點意見!!  一步一步慢慢地朝向本土化自給自足邁進!!

The figures here are being typeset into this manuscript and this manuscript, using Xelatex with TeX life 2013. And because my document is latex, I saved the above figures in PGF format using Python pgf backend support (\usepackage{pgf}). I found it very useful. So truly thanks to those who contributed to the pgf latex support in Python!

Some hindsight (still amending): I used to use the software Origin and Matlab to draw my research data on charts. But they are not free, and are actually somehow expensive. After I learned a language Python I realized, hmm, perhaps what we learned in school, like Matlab or MS words, powerpoint, etc, are probably not the best choices to learn during education. Although a programming language is not easy to learn, I somehow feel it is worth while to learn because we will need to make graphs and create text documents and maybe create professional typeset documents our whole life. A programming language is not meant to learn in 1 or 2 years, but learning is a life-long process, so there is no need to learn it in just 1 or 2 years. We can extend the learning period to 5 to 10 years, and still have plenty of time to apply it because we will need to deal with documents all our life.

---------------------------------------

Some open issues with mplot3d in Matplotlib, Python:
2014/7/21

Can not set equal aspect ratio on 3D axes. This has to be done by manually creating a square figure and a square invisible cube. 

Perspective/orthogonal effect is not yet integrated. A workaround is provided here. It works but it will generate errors when cursor hover over the figure. Also, once this workaround is set, not sure how to set it back (or turn it back off). 

Seems its still impossible to change the LineCollection position data in a 3D animation. 

Saving as eps files.[1][2][3]

Keywords: 3D CAD, 3D vector graphs

推薦延伸閱讀：
用python繪製3d內切內接圓
http://whymranderson.blogspot.tw/2015/04/tetrahedron-midsphere.html