CSCI 4152/6509 Natural Language Processing. Lab 1: FCS Computing Environment

Transcription

1 CSCI 4152/6509 Natural Language Processing Lab 1: FCS Computing Environment vlado/csci6509 Lab Instructor: Dijana Kosmajac, Dhivya Jayaraman Slides copyright: Mike McAllister, Vlado Keselj, Magdalena Jankowska Faculty of Computer Science Dalhousie University 12-Jan-2018 (1) CSCI 4152/6509 1

2 Lab Overview An objective: Make sure that all students are familiar with their CSID and how to login to the bluenose server Refresh your memory about Linux Command-Line Interface Introduction to Perl (just hello world for now) Introduction to your SVN repository for the course Note 1: Replace csuserid with your FCS CS user id (a.k.a., CSID) Note 2: If you do not know your CSID, lookup at: 12-Jan-2018 (1) CSCI 4152/6509 2

3 Step 1: Logging in to server bluenose You can choose Windows, Mac or Linux environment in some labs Windows: you will use PuTTY program On Mac: open a Terminal and type: ssh csuserid@bluenose.cs.dal.ca (instead of csuserid use your CS userid ) On Linux: similarly to Mac, you open the terminal and type the same command: ssh csuserid@bluenose.cs.dal.ca 12-Jan-2018 (1) CSCI 4152/6509 3

4 Running PuTTY If you use Windows, then one option is to use PuTTY to login to the bluenose server Double-click the PuTTY icon, and the following window should appear: 12-Jan-2018 (1) CSCI 4152/6509 4

5 12-Jan-2018 (1) CSCI 4152/6509 5

6 Review of Some Linux Commands Step 2: pwd Step 3: mkdir csci4152 or mkdir csci6509 ls chmod go-rx csci6509 or chmod go-rx csci4152 Step 4: cd csci6509 or cd csci4152 Make directory lab1 and change your current directory to it. 12-Jan-2018 (1) CSCI 4152/6509 6

7 Step 5: Using emacs prepare hello.pl Use an editor: emacs or some other (e.g., vi, pico... ) #!/usr/bin/perl print "Hello world!\n"; Step 6: Running a Perl program perl hello.pl Another way to run the program: chmod u+x hello.pl./hello.pl 12-Jan-2018 (1) CSCI 4152/6509 7

8 Source Control with SVN In this tutorial, you will learn or refresh your knowledge about SVN. Performing these steps will be graded as a part of an assignment SVN will be used for assignment submission It is possible that not all students have the course SVN repository. In that case, let the instructor know. You will be able to finish the tutorial next week. This tutorial is slightly modified version of the SVN tutorial Source Control with SVN prepared by Michael McAllister for CSCI Jan-2018 (1) CSCI 4152/6509 8

9 Source Control with SVN Subversion (or SVN) source control system SVN can be accessed using command line interface on bluenose (Linux system); similarly on Mac GUI interfaces exist (including Windows) Development companies use some form of source control SVN is relatively popular, open-source system git is another newer open-source system (e.g., used for Linux kernel) Some commercial software: Perforce and ClearCase Predecessors: SCCS, RCS, CVS 12-Jan-2018 (1) CSCI 4152/6509 9

10 1. checkout projects in SVN Objectives 2. add a subproject (subdirectory) to SVN 3. add, delete files in SVN; and commit your changes 4. change files in SVN, and merge changes from different copies 5. resolve conflicts that may result from parallel changes 6. revert changes and retrieve older versions of the files 12-Jan-2018 (1) CSCI 4152/

11 Repository and Working Copy SVN uses remote directory system called repository to store files and history of their changes You can review changes and retrieve older versions Files can be shared among users and SVN can be used for collaborative work SVN uses notion of working copy of the files We send (or commit) files to the repository when we are happy with their state For example, we generally do not commit program unless they can be compiled 12-Jan-2018 (1) CSCI 4152/

12 A Couple Notes Replace csuserid with your CS user id (a.k.a., CSID) Important Security Note: Each time that you enter an SVN command in this lab, you will be asked for your bluenose password. Enter it. When asked if you want to store the password, answer no each time, since the password is not stored safely currently. If you are sure that you know what you are doing, you may temporarily say yes to make SVN use easier, but then make sure that only you have access permissions to the the directory./subversion, and it is a good idea to delete the whole directory after each session. 12-Jan-2018 (1) CSCI 4152/

13 Step 1: Set-up Before SVN Tutorial Log into bluenose in two different windows (use putty twice (windows) or ssh from two terminals (mac os)). Call one window your primary window and the other window your secondary window. In your course directory in subdirectory lab1, create two subdirectories, one called primary and another called secondary. In your primary window, make primary the working directory. In your secondary window, make secondary the working directory. 12-Jan-2018 (1) CSCI 4152/

14 Step 2: Checkout and Adding a Directory We use checkout command to retrieve a working copy In your primary window run (in one line): svn checkout You will be asked to enter the password Change directory using: cd csuserid List directory using: ls -a Notice.svn directory Create directory: mkdir svnlab Add it to the repository: svn add svnlab You should see output: A svnlab 12-Jan-2018 (1) CSCI 4152/

15 Step 3: Commit Command The svnlab directory is not in the repository yet We must execute commit to send it to the repository (all in one line): svn commit -m "Creating a project for my SVN lab" You should see the output: Adding svnlab Committed revision 2. Check the following URL using your browser: 12-Jan-2018 (1) CSCI 4152/

16 SVN Log Messages -m "xx" part adds a message to the repository All commit commands require a log message 12-Jan-2018 (1) CSCI 4152/

17 Step 4.a: Adding Files: Preparing Files Create a file name hello.pl with the following content: #!/usr/bin/perl print "Hello world!\n"; Create two copies of this file: cp hello.pl hello2.pl cp hello.pl tmp.pl 12-Jan-2018 (1) CSCI 4152/

18 Step 4.b: Adding Files Add two files: svn add hello.pl hello2.pl We see the output: A A hello.pl hello2.pl Run command: svn status Output: A hello.pl A hello2.pl? tmp.pl 12-Jan-2018 (1) CSCI 4152/

19 Step 4.c: Adding Files: Commit Commit files (one line): svn commit -m "Adding the first version of my sample files." Enter: svn status Output:? tmp.pl 12-Jan-2018 (1) CSCI 4152/

20 Step 5: Deleting Files (5.a) Delete a file: svn delete hello2.pl Output: D hello2.pl Enter: svn status Output:? tmp.pl D hello2.pl 12-Jan-2018 (1) CSCI 4152/

21 Step 5.b: Deleting a File (2) Enter in one line: svn commit -m "File hello2.pl was added by mistake earlier. It was just a backup copy of hello.pl" The file is deleted, but it can be retrieved from the repository 12-Jan-2018 (1) CSCI 4152/

22 Step 6: Checking out Second Copy Go the secondary window and make sure you are in the directory:... /lab1/secondary/ You can check this using command: pwd In secondary window get the second working copy: svn checkout nlp-course/csuserid/svnlab Change directory: cd svnlab List the contents: ls -la 12-Jan-2018 (1) CSCI 4152/

23 Step 7: Changing Files (7.a) In your primary window edit hello.pl to become the following file: #!/usr/bin/perl for my $i (1..10) { print "Hello world!\n"; } Run: svn commit In the secondary window run: svn update Verify that the file has been updated 12-Jan-2018 (1) CSCI 4152/

24 7.b: Modifying a File In your primary window: edit your copy of hello.pl and change the for loop to iterate 15 times rather than 10 tiems, and save the file. Run: svn status You should get the following output: M hello.pl? tmp.pl M indicates that the file has some modifications which are not in the repository 12-Jan-2018 (1) CSCI 4152/

25 Step 7.c: svn diff command Run: svn diff hello.pl Output will show differences between working version and repository version Commit the change: svn commit -m "Use a longer loop iteration" Try again svn diff hello.pl No output. Why? In the secondary window look at hello.pl 12-Jan-2018 (1) CSCI 4152/

26 Step 7.d: Update In the secondary window run: svn update Output: Updating. : U hello.pl Updated to revision 36. U means that there is a more recent copy in the repository Check again file hello.pl 12-Jan-2018 (1) CSCI 4152/

27 Step 7.e: Delete Let us remove a file: rm hello.pl Check that it does not exist: ls (Still in secondary window) Run: svn status Output:! hello.pl! is an indicator that the file is missing Run: svn update The file hello.pl is restored 12-Jan-2018 (1) CSCI 4152/

28 Step 7.f: Parallel Changes Let us change hello.pl in both windows: In primary window add a print statement before the loop: print "I am about to start the loop\n"; In secondary window add a print statement after the loop: print "I have finished the loop\n"; In each windows use svn status to verify that the working file is different from the repository copy (indicator M) 12-Jan-2018 (1) CSCI 4152/

29 Step 7.g: Merging Changes In primary window commit changes, e.g.: svn commit -mprimary Try to commit changes in secondary window, e.g.: svn commit -msecondary Commit fails since the repository copy changed since our last update Run: svn update Output: G hello.pl Updated to revision 37. G indicates that changes have been merged 12-Jan-2018 (1) CSCI 4152/

30 Step 7.g (cont): Merging Changes (Still in secondary window) Edit file hello.pl to make sure you are happy with merged changes Run: svn commit -m"merged changes" This will submit merged changes to repository In both windows run: svn update Now both windows contain the latest copy in repository 12-Jan-2018 (1) CSCI 4152/

31 Step 7.h: Conflicting Changes In primary window change hello.pl so that the loop iterates 30 times Commit the change In secondary window change hello.pl so that the loop iterates 5 times Try to commit the change Similarly as before commit fails, so we need to do update first Run svn update However, changes are not merged this time 12-Jan-2018 (1) CSCI 4152/

32 We get the following message Conflict discovered in hello.pl. Select: (p) postpone, (df) diff-full, (e) edit, (mc) mine-conflict, (tc) theirs-conflict, (s) show all options: df see differences, or different options of editing or fixing the file Let us choose to postpone, enter: p We can see now several files in secondary window by running: ls The files are (likely with different version numbers): hello.pl the hello.pl file with the diff information embedded hello.pl.mine your modified code hello.pl.r38 the original copy of the code that you modified in this working directory (revision 38) hello.pl.r39 the most recent version of the code in the repository (revision 39) 12-Jan-2018 (1) CSCI 4152/

33 Step 7.i: Examining Conflict (Still in secondary window) Run status command Output will include line: C hello.pl C indicates conflict that needs to be resolved Examine the log using log command: svn log hello.pl The output shows log messages (which demonstrate why it is useful to write descriptive log messages) Additional information can be obtained using diff command: svn diff hello.pl A part of the output is: +<<<<<<<.mine +for my $i (1..5) { +======= for my $i (1..30) { +>>>>>>>.r39 12-Jan-2018 (1) CSCI 4152/

34 Step 7.j: Resolving Conflict Fix hello.pl file. Usually, you can either: Edit hello.pl directly (delete extra lines, etc.) Choose our version: cp hello.pl.mine hello.pl Choose repository version: cp hello.pl.r39 hello.pl Choose original last updated version: cp hello.pl.r38 hello.pl We can simply choose: cp hello.pl.mine hello.pl To indicate that the conflict has been resolved, and hello.pl is the version that we want submitted to the repository, we issue svn resolved command: svn resolved hello.pl Other files (hello.pl.mine and hello.pl.rx) are removed Finally, we commit the change: svn commit -m"conflict resolved" 12-Jan-2018 (1) CSCI 4152/

35 Step 8: Retrieving copies of previous work 8.a: Sometimes we make changes to a program and want to undo them It is useful to make commit each time we reach a stable point Aside: If collaborating with someone, it is useful to do frequent updates and commits In secondary window: Introduce error in file hello.pl; e.g., change word print to pprint Save the file but do not commit Suppose that we want to undo the change to the last saved copy in the repository. We already know three options: make checkout elsewhere and get the fresh copy from there delete hello.pl and run update use diff to identify changes and manually reverse them All these involve unnecessary work; much easier: svn revert hello.pl 12-Jan-2018 (1) CSCI 4152/

36 Step 8.b: Retrieving Older Versions How to retrieve older versions of a file? We can use -r option with the update command: svn update -r 34 hello.pl We can use log command (svn log hello.pl) to find interesting revision numbers Check the contents of hello.pl If you change this file and try to commit, it will not work as a newer version is in the repository. You would need to make an update, and likely resolve any conflicts. Instead, simply run: svn update This will get the latest version from the repository 12-Jan-2018 (1) CSCI 4152/

37 Step 8.c: Retrieving by Timestamps Sometimes we do not know exact revision number, but we know time when the directory was last stable. We can specify time to retrieve the latest version that existed at this time For example: svn update -r "{ :00:00}" 12-Jan-2018 (1) CSCI 4152/

38 Step 9: End of the Lab After finishing with commits, you can even delete primary and the secondary directory, since all required updates are submitted to the SVN. If you want, you can still keep them until Assignment 1 is marked, just to be sure. To delete the directories, you will need to be in their parent directory, and use the command rm -r primary secondary End of the lab. 12-Jan-2018 (1) CSCI 4152/