-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathlog2014_06_19.html
165 lines (165 loc) · 15.2 KB
/
log2014_06_19.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
<!DOCTYPE html>
<html lang='en'>
<head>
<title>
pgmpy Log of 19-06-2014
</title>
<link rel='stylesheet' type='text/css' href='css/bootstrap.min.css' media='screen' />
</head>
<body>
<div class='container'>
<h2>Log of the <code>#pgmpy</code> IRC Channel</h2><br/>
<h3>All the times shown here presently are in Indian Standard Time(IST) +0530Hrs<h3/>
<h3>Date : 19-06-2014 </h3><br/><br/>
<br/>
<kbd>[06:43:11]</kbd> <strong>[connected at Thu Jun 19 06:43:11 2014]</strong><br/>
<kbd>[06:43:26]</kbd> <strong>[I have joined #pgmpy]</strong><br/>
<kbd>[06:43:28]</kbd> <code><ankurankan></code> Great!!<br/>
<kbd>[06:43:37]</kbd> <code><debangshu></code> but what is the func of analysis mode<br/>
<kbd>[06:43:55]</kbd> <code><ankurankan></code> We will be using analysis mode to run some algos on the graph<br/>
<kbd>[06:44:29]</kbd> <code><ankurankan></code> But we are focusing on draw mode right now<br/>
<kbd>[06:44:41]</kbd> <code><ankurankan></code> Because algo are still not implemented in the pgmpy<br/>
<kbd>[06:45:10]</kbd> <code><debangshu></code> ok<br/>
<kbd>[06:45:41]</kbd> <code><debangshu></code> pgmpy_viz is running without pgmpy<br/>
<kbd>[06:46:00]</kbd> <code><ankurankan></code> Yeah right now its not connected with the backend<br/>
<kbd>[06:46:09]</kbd> <code><ankurankan></code> Its just JS that's working now<br/>
<kbd>[06:46:38]</kbd> <code><debangshu></code> yes..i checked.<br/>
<kbd>[06:47:22]</kbd> <code><debangshu></code> backend will provide RESTful end points.i guess?<br/>
<kbd>[06:47:22]</kbd> <code><ankurankan></code> So just try to play with things and tell us something that you think can be improved<br/>
<kbd>[06:47:34]</kbd> <code><ankurankan></code> Yeah<br/>
<kbd>[06:49:40]</kbd> <code><debangshu></code> ok. and about a node. i think some probability parameter is needed to be stored<br/>
<kbd>[06:50:29]</kbd> <code><ankurankan></code> Yeah<br/>
<kbd>[06:50:50]</kbd> <code><ankurankan></code> Basically it won't be just one probability value<br/>
<kbd>[06:50:58]</kbd> <code><ankurankan></code> It will be a table<br/>
<kbd>[06:51:20]</kbd> <code><ankurankan></code> And this table size would be variable for each node<br/>
<kbd>[06:52:19]</kbd> <code><ankurankan></code> we will need to compute the size of the table in the frontend only<br/>
<kbd>[06:54:06]</kbd> <code><debangshu></code> ok<br/>
<kbd>[06:55:03]</kbd> <code><debangshu></code> tell me something about the table. Is it about storing probability w.r.t. other nodes<br/>
<kbd>[06:55:05]</kbd> <code><debangshu></code> ?<br/>
<kbd>[06:57:25]</kbd> <code><ankurankan></code> Yeah. If you saw PGM, each node has a Conditional Probability Distribution (CPD) which is a table of probabilities<br/>
<kbd>[06:58:07]</kbd> <code><ankurankan></code> And this table size depends on the parents of the node<br/>
<kbd>[07:58:33]</kbd> <code><debangshu></code> i have seen an issue stating size of cy should not be fixed.<br/>
<kbd>[07:59:08]</kbd> <code><debangshu></code> then why we are not using col-md-12 etc as in bootstrap?<br/>
<kbd>[17:19:20]</kbd> <code><shashankg></code> harsh11: Hi<br/>
<kbd>[17:19:41]</kbd> <code><harsh11></code> shashankg :hi<br/>
<kbd>[17:20:06]</kbd> <code><shashankg></code> Let's wait for ankur, then we'll commence the meeting.<br/>
<kbd>[17:20:19]</kbd> <code><harsh11></code> sure<br/>
<kbd>[17:20:23]</kbd> <code><shashankg></code> Till then let's introduce ourselves.<br/>
<kbd>[17:20:41]</kbd> <code><harsh11></code> ok . I am Harsh Rawat<br/>
<kbd>[17:21:07]</kbd> <code><harsh11></code> I am a student in ECE at IIT(BHU)<br/>
<kbd>[17:21:50]</kbd> <code><shashankg></code> cool.<br/>
<kbd>[17:22:11]</kbd> <code><shashankg></code> I am Shashank Garg, I was a student of ECE IIT(BHU)<br/>
<kbd>[17:22:30]</kbd> <code><shashankg></code> You are from which year?<br/>
<kbd>[17:22:48]</kbd> <code><harsh11></code> i am a 3rd yearite<br/>
<kbd>[17:23:45]</kbd> <code><shashankg></code> Okay<br/>
<kbd>[17:24:59]</kbd> <code><shashankg></code> have you been through our github repo and the document I linked on Facebook comment?<br/>
<kbd>[17:25:58]</kbd> <code><harsh11></code> yeah i saw through it superficially.<br/>
<kbd>[17:27:20]</kbd> <code><shashankg></code> would you like to work in a python+js task or js only task.<br/>
<kbd>[17:28:54]</kbd> <code><harsh11></code> python +js task would be better .... Though i have somewhat less knowledge of js..<br/>
<kbd>[17:30:27]</kbd> <code><shashankg></code> ok. The first two tasks are mainly python. JS would be required at only some points.<br/>
<kbd>[17:31:19]</kbd> <code><harsh11></code> yes that would be better as i can pick up js in that time .<br/>
<kbd>[17:34:40]</kbd> <code><shashankg></code> Ankur will not be able to join us today.<br/>
<kbd>[17:34:47]</kbd> <code><shashankg></code> Let's continue on our own.<br/>
<kbd>[17:35:09]</kbd> <code><harsh11></code> Sure<br/>
<kbd>[17:35:24]</kbd> <code><shashankg></code> I think you don't have any knowledge of pgm<br/>
<kbd>[17:36:06]</kbd> <code><shashankg></code> Probabilistic graphical models<br/>
<kbd>[17:36:06]</kbd> <code><harsh11></code> Yes i dont have knowledge of pgm right now.<br/>
<kbd>[17:36:31]</kbd> <code><shashankg></code> It's ok, you won't be needing any knowledge for now.<br/>
<kbd>[17:37:03]</kbd> <code><harsh11></code> ok.. so what would be the priliminary work<br/>
<kbd>[17:37:37]</kbd> <code><shashankg></code> The definition will suffice for now. Basically pgm represents the probabilistic relationship between some variables using graphs.<br/>
<kbd>[17:38:21]</kbd> <code><harsh11></code> for pgm i had downloaded coursera videos.<br/>
<kbd>[17:38:47]</kbd> <code><harsh11></code> Will they be suffice for the project?<br/>
<kbd>[17:38:53]</kbd> <code><shashankg></code> ok. They are a great source. You can also read the book by koller.<br/>
<kbd>[17:39:31]</kbd> <code><shashankg></code> Yeah, they will suffice. For the initial few tasks you won't be needing any pgm.<br/>
<kbd>[17:39:59]</kbd> <code><shashankg></code> pgmpy currently has two parts. pgmpy and pgmpy_viz.<br/>
<kbd>[17:40:27]</kbd> <code><harsh11></code> Yes ankur bhaiya mentioned it<br/>
<kbd>[17:40:54]</kbd> <code><shashankg></code> pgmpy models the graphs and implements the pgm algorithms on these graphs. viz works for vizualisation in a browser.<br/>
<kbd>[17:41:08]</kbd> <code><shashankg></code> Right now we are working on integrating them.<br/>
<kbd>[17:41:32]</kbd> <code><shashankg></code> We are using json to communicate with the browser.<br/>
<kbd>[17:41:52]</kbd> <code><harsh11></code> ok.<br/>
<kbd>[17:42:23]</kbd> <code><shashankg></code> Since python is your forte, I'll explain the first two tasks to you, you can pick any of them.<br/>
<kbd>[17:42:31]</kbd> <code><shashankg></code> task#1<br/>
<kbd>[17:43:13]</kbd> <code><shashankg></code> There are some standard file formats for representing a graphical mode.<br/>
<kbd>[17:43:20]</kbd> <code><harsh11></code> yes<br/>
<kbd>[17:43:32]</kbd> <code><shashankg></code> The list of these models is in the task doc.<br/>
<kbd>[17:44:07]</kbd> <code><shashankg></code> So what you need to do is, you have to accept the file from user, in a browser.<br/>
<kbd>[17:44:07]</kbd> <code><harsh11></code> yes<br/>
<kbd>[17:45:18]</kbd> <code><shashankg></code> I think the it would be fairly easy to create a uploading framework.<br/>
<kbd>[17:45:48]</kbd> <code><shashankg></code> The library already has code to parse these file formats.<br/>
<kbd>[17:46:01]</kbd> <code><harsh11></code> yes i would be able to do that<br/>
<kbd>[17:46:43]</kbd> <code><shashankg></code> You have to parse the file, you got from browser and from the retrieved information you have to create a json<br/>
<kbd>[17:47:01]</kbd> <code><shashankg></code> The format of json is also defined, and linked in task doc.<br/>
<kbd>[17:47:16]</kbd> <code><harsh11></code> ok<br/>
<kbd>[17:47:51]</kbd> <code><shashankg></code> Now JS comes into the scene.<br/>
<kbd>[17:48:36]</kbd> <code><shashankg></code> You have to draw the graph using JS and cytoscape in the browser using the information in json.<br/>
<kbd>[17:49:28]</kbd> <code><harsh11></code> ok .<br/>
<kbd>[17:49:52]</kbd> <code><shashankg></code> Figuring out, how to create the graph will not be tricky.<br/>
<kbd>[17:50:28]</kbd> <code><shashankg></code> viz already has the graph drawing functionality, all you have to do is automate the process for your json.<br/>
<kbd>[17:50:30]</kbd> <code><harsh11></code> yes, i libraries can be used for that<br/>
<kbd>[17:50:45]</kbd> <code><shashankg></code> So how do you find the task?<br/>
<kbd>[17:51:32]</kbd> <code><harsh11></code> Its good and i think i can complete it with some reading<br/>
<kbd>[17:51:42]</kbd> <code><shashankg></code> cool :)<br/>
<kbd>[17:52:11]</kbd> <code><shashankg></code> Now task#2<br/>
<kbd>[17:52:15]</kbd> <code><harsh11></code> task<br/>
<kbd>[17:53:05]</kbd> <code><shashankg></code> user can manually draw graphs in browser using viz.<br/>
<kbd>[17:53:41]</kbd> <code><shashankg></code> When the user creates the graph, the graph information is stored in a json and then sent to backend.<br/>
<kbd>[17:54:41]</kbd> <code><shashankg></code> what you have to do is: you have to send the json from browser, and accept it at backend<br/>
<kbd>[17:55:05]</kbd> <code><harsh11></code> ok<br/>
<kbd>[17:55:10]</kbd> <code><shashankg></code> now you have to parse the graph using python, and create a graph in pgmpy.<br/>
<kbd>[17:56:00]</kbd> <code><shashankg></code> pgmpy already has functions for creating the graph, you just have to figure out certain library functions and use them for it.<br/>
<kbd>[17:56:09]</kbd> <code><harsh11></code> ok ..<br/>
<kbd>[17:56:42]</kbd> <code><shashankg></code> Do you have any knowledge of git or any other vcs(version control system)?<br/>
<kbd>[17:57:15]</kbd> <code><shashankg></code> ankurankan: Hi<br/>
<kbd>[17:57:41]</kbd> <code><harsh11></code> Nope. Ithis is my first time with Open SOurce and related tools<br/>
<kbd>[17:58:32]</kbd> <code><shashankg></code> ok.<br/>
<kbd>[17:58:46]</kbd> <code><shashankg></code> git is very east to learn.<br/>
<kbd>[17:59:13]</kbd> <code><shashankg></code> So for starting, create a github account.<br/>
<kbd>[17:59:37]</kbd> <code><shashankg></code> fork and clone the repo.<br/>
<kbd>[17:59:46]</kbd> <code><harsh11></code> Done that<br/>
<kbd>[18:00:02]</kbd> <code><shashankg></code> cool<br/>
<kbd>[18:00:41]</kbd> <code><shashankg></code> first three chapters of progit will suffice for all the git we need.<br/>
<kbd>[18:00:57]</kbd> <code><harsh11></code> ok<br/>
<kbd>[18:01:39]</kbd> <code><shashankg></code> which task would you prefer for the beginning?<br/>
<kbd>[18:02:01]</kbd> <code><shashankg></code> you can also decide later after going through the code<br/>
<kbd>[18:02:02]</kbd> <code><harsh11></code> First task would be better suited to me.<br/>
<kbd>[18:02:06]</kbd> <code><shashankg></code> ok<br/>
<kbd>[18:03:25]</kbd> <code><shashankg></code> If you want to go through this conversation again<br/>
<kbd>[18:03:25]</kbd> <code><harsh11></code> :)<br/>
<kbd>[18:03:40]</kbd> <code><shashankg></code> you can read the logs at bit.ly/1jLjsHf<br/>
<kbd>[18:04:11]</kbd> <code><harsh11></code> Ok .. I understood the work to be done<br/>
<kbd>[18:04:21]</kbd> <code><shashankg></code> If you have any questions, you can ask<br/>
<kbd>[18:05:10]</kbd> <code><harsh11></code> Can i have your e-mail address so that i can ask any doubts later<br/>
<kbd>[18:05:50]</kbd> <code><shashankg></code> If you have any doubts you can mail them at [email protected]<br/>
<kbd>[18:06:28]</kbd> <code><harsh11></code> ok<br/>
<kbd>[18:06:44]</kbd> <code><shashankg></code> my mail id is [email protected] but it would be better if you mail it on group, If I am busy someone else will answer you doubt..<br/>
<kbd>[18:07:02]</kbd> <code><ankurankan_></code> harsh11, Also you can join this irc channel any time<br/>
<kbd>[18:07:11]</kbd> <code><ankurankan_></code> I am mostly available here<br/>
<kbd>[18:07:32]</kbd> <code><harsh11></code> ok.Will do if i have to ask something<br/>
<kbd>[18:07:58]</kbd> <code><harsh11></code> What would be the deadline for the work?<br/>
<kbd>[18:08:19]</kbd> <code><ankurankan_></code> harsh11, This is open source. There are no deadlines :)<br/>
<kbd>[18:08:26]</kbd> <code><harsh11></code> :)<br/>
<kbd>[18:08:50]</kbd> <code><ankurankan_></code> harsh11, Work whenever you are free.<br/>
<kbd>[18:09:00]</kbd> <code><ankurankan_></code> harsh11, But if you are working on a feature try to complete it as soon as possible<br/>
<kbd>[18:09:01]</kbd> <code><harsh11></code> Ok. I'll try to complete this work soon<br/>
<kbd>[18:09:06]</kbd> <code><shashankg></code> read the docs of the project at http://pgmpy.readthedocs.org/en/latest/ for better understanding.<br/>
<kbd>[18:09:21]</kbd> <code><ankurankan_></code> harsh11, So that if you couldn't do it, someone else will try.<br/>
<kbd>[18:09:44]</kbd> <code><harsh11></code> Yes, I understand.<br/>
<kbd>[18:10:13]</kbd> <code><shashankg></code> And go through the wikis on github. They will help you with git and some other important issues.<br/>
<kbd>[18:10:58]</kbd> <code><harsh11></code> Yes will do.<br/>
<kbd>[18:11:00]</kbd> <code><ankurankan_></code> harsh11, And if you get stuck anywhere just ask it<br/>
<kbd>[18:11:02]</kbd> <code><shashankg></code> https://github.com/pgmpy/pgmpy_viz/wiki<br/>
<kbd>[18:12:05]</kbd> <code><shashankg></code> So, I think we are done for today.<br/>
<kbd>[18:12:26]</kbd> <code><harsh11></code> Ok. I'll be in touch.<br/>
<kbd>[18:12:36]</kbd> <code><ankurankan_></code> shashankg, Didn't anyone else join the meeting today?<br/>
<kbd>[18:12:50]</kbd> <code><shashankg></code> No<br/>
<kbd>[18:13:12]</kbd> <code><ankurankan_></code> Okay..<br/>
<kbd>[18:13:18]</kbd> <code><ankurankan_></code> I will try to contact them<br/>
<kbd>[18:14:21]</kbd> <code><shashankg></code> ok. Tell them to read the logs. We can also have a meeting for new people.<br/>
<kbd>[18:14:29]</kbd> <code><ankurankan_></code> okay..<br/>
<kbd>[14:29:08]</kbd> <strong>[disconnected at Sat Jun 21 14:29:08 2014]</strong><br/>
</div>
<script src='https://ajax.googleapis.com/ajax/libs/jquery/1.11.0/jquery.min.js'></script>
<script src='//netdna.bootstrapcdn.com/bootstrap/3.1.1/js/bootstrap.min.js'>
</script>
</body>
</html><br/>