The bug in debugging a function is no longer an issue in python 3. Should the section be removed entirely, or should a new bug be created, so students can continue to experience the joys of debugging?

One possibility would be to introduce an order of operations bug, temp - 32 * 5/9 instead of (temp - 32) * (5/9).

In lesson 01-numpy and 04-files, we are teaching that matplotlib.pyplot.show take in arguments either graphics or figures. For example:

ave_inflammation = data.mean(axis=0)
ave_plot = matplotlib.pyplot.plot(ave_inflammation)
matplotlib.pyplot.show(ave_plot)

If we refer to matplotlib API, there are no mentions that show takes matplotlib object as arguments. The stated behaviour is that show will display all figures and blocks.

I agree that it would be useful to be able to tell matplotlib which graphic we want to display, but the API as it is now does not support it. To convince ourselves, simply run the following code:

import numpy

data = numpy.loadtxt(fname='inflammation-01.csv', delimiter=',')

import matplotlib.pyplot

ave_inflammation = data.mean(axis=0)
ave_plot = matplotlib.pyplot.plot(ave_inflammation)
max_plot = matplotlib.pyplot.plot(data.max(axis=0))
min_plot = matplotlib.pyplot.plot(data.min(axis=0))
matplotlib.pyplot.show(ave_plot)

What we would expect is to only see the average plot, while we end up with all three graphics on the same figure.

From my point of view, since the arguments pass to show are simply drop, it does not serve any purpose to tell students to pass graphics or figure to show, and should simply call show without arguments in the courses.

From https://github.com/swcarpentry/lesson-template/blob/gh-pages/tools/check.py:

• Heading at line 41 should be level 2
• Heading at line 433 should be level 2
• Heading at line 500 should be level 2
• Heading at line 524 should be level 2
• Heading at line 551 should be level 2
• Heading at line 574 should be level 2
• Heading at line 605 should be level 2
• Document contains heading not specified in the template: Assertions
• Document contains heading not specified in the template: Test-Driven Development
• Document contains heading not specified in the template: Debugging
• Document contains heading not specified in the template: Know What It's Supposed to Do
• Document contains heading not specified in the template: Make It Fail Every Time
• Document contains heading not specified in the template: Make It Fail Fast
• Document contains heading not specified in the template: Change One Thing at a Time, For a Reason
• Document contains heading not specified in the template: Keep Track of What You've Done
• Document contains heading not specified in the template: Be Humble
• Could not find the linked asset file ../../rules.html in /tmp/python/../../rules.html. If this is a URL, it must be prefixed with http(s):// or ftp://.
• The topic page should not have sub-headings outside of special blocks. If a topic needs sub-headings, it should be broken into multiple topics.

It would be nice to have the different axes labeled patients and days in this diagram.

https://github.com/swcarpentry/python-novice-inflammation/blob/gh-pages/fig/python-operations-across-axes.svg

Since Pandoc can generate figure captions for images, it would be nice to be able to use figure numbering to refer to figures from the text (Figure 1, Figure 2, etc.). Currently, the figure captions are escaped by using the <backslash><white-space> syntax.

This is currently not possible directly in Pandoc, but there are some solutions discussed by @r-gaia-cs that involve CSS or Pandoc filters.

Every time I read the command line lesson (08-cmdline.html), the files readings-01.py, readings-02.py, etc make me very uncomfortable, because that type of file naming behavior is exactly what we are trying to avoid by teaching version control.

I think the solution to this problem comes from the very beginning of the lesson, which reads:

... save the following in a text file called sys-version.py:

import sys
print 'version is', sys.version

We could take a similar approach to the rest of the lesson and instead of saying $ cat readings-01.py we could simply save the following in a text file called readings.py:

import sys
import numpy as np

def main():
    script = sys.argv[0]
    filename = sys.argv[1]
    data = np.loadtxt(filename, delimiter=',')
    for m in data.mean(axis=1):
        print m

Later in the lesson when we get to readings-02.py, we could just say that we open our text editor and make the following changes to readings.py:

import sys
import numpy as np

def main():
    script = sys.argv[0]
    filename = sys.argv[1]
    data = np.loadtxt(filename, delimiter=',')
    for m in data.mean(axis=1):
        print m

main()

The trick would be to somehow highlight main() in bold or a different color, because that is the part of the script that has changed since last time (i.e. kind of like git diff highlighting).

If people are happy with this suggestion of altering the wording of the command line lesson to make successive edits to readings.py as opposed to going $ cat readings-01.py, $ cat readings-02.py etc then I'd be happy to submit a pull request. (I'm also open to suggestions of how to incorporate git diff style highlighting into that PR, but I'd have to be able to write it in markdown and then and have pandoc convert it appropriately - I'm not aware of any way to make that happen.)

Code like weight_lb = 2.2 * weight_kg should be weight_lb = 2.2*weight_kg according to PEP8. Does this matter here?

Currently, we teach TDD using range_overlap as an example. I have found that the algorithm is not immediately obvious to learners, contributing significant extraneous load. I'd like to know if other instructors have had this problem and a mind to change it in our lessons. I would propose something much simpler, like

def common(first, second):
    """ Returns a (sorted) list of elements common to the
        lists 'first' and 'second'
    """

I'm happy to implement this along with some tests if there is any interest.

Please use this issue for discussion of changes we should make by mid-August 2015.

I'm not sure of the intention of the html version, but wouldn't it make sense include a direct link to the csv data files? That way students can work on this tutorial at their own pace without having to clone the repository?
See also #9.

See for example at Errors and Exceptions:

http://swcarpentry.github.io/python-novice-inflammation/instructors.html

From https://github.com/swcarpentry/lesson-template/blob/gh-pages/tools/check.py:

• Heading at line 61 should be level 2
• Document contains heading not specified in the template: Debugging a Function
• Document contains heading not specified in the template: Composing Functions
• Document contains heading not specified in the template: The Call Stack
• Document contains heading not specified in the template: Testing and Documenting
• Document contains heading not specified in the template: Defining Defaults
• Could not find the linked asset file 01-numpy.ipynb in /tmp/python/01-numpy.ipynb. If this is a URL, it must be prefixed with http(s):// or ftp://.
• The topic page should not have sub-headings outside of special blocks. If a topic needs sub-headings, it should be broken into multiple topics.

The code and the data used in the lesson 06-cmdline have been moved to separate subdirectories. Thus the code as written in the lesson would not be executable. Is your plan to have this layout for the repo only, and then in live workshops have everything in one directory?

I ask because I know that novices struggle with navigating directories (most will have just seen it for the first time in the shell lesson). For the novice R lessons, we are trying to keep the source executable in R Markdown files. For this to work, what is written in the lesson has to correspond to the actual layout in the lesson repository. We recently moved the inflammation files to a subdirectory and updated the lessons. However, we have not yet moved the scripts. I hesitate to do so because this lesson is already challenging, and it is inevitable that many students will save the R script in the current working directory because they did not cd code before running nano.

I'd appreciate any advice on how to proceed.

From an email I received:

Dear Software Carpentry,

Thank you for doing this work. I refer all scientists I know who express interest in becoming better programmers to your website and lessons.

I have one small bug report. I was reading your Python lesson on conditionals. One of the learning objectives is explaining the difference between tuples and lists, but tuples are not mentioned again in the lesson.
http://swcarpentry.github.io/python-novice-inflammation/05-cond.html

Also tuples and ipythonblocks library are mentioned in the reference material, but are not present in the lesson.
http://swcarpentry.github.io/python-novice-inflammation/reference.html

Thanks again for making this service available.

The "Tabs and Spaces" callout in lesson 7 includes a code example where supposedly "the first two lines are using a tab for indentation, while the third line uses four spaces". In fact, all three lines use spaces for indentation.

Someone copying this example would not be able to reproduce the error.

I'm opening an issue rather than just fixing it, because:

1. this might be a deliberate decision (the markdown file also has spaces, so this is not just a conversion problem)?
2. fixing it is non-trivial, given that pandoc converts tabs to spaces by default (the '--preserve-tabs' option would have to be added to the pandoc call in the Makefile).

Just wondering what the data is a measurement of. It is certainly not temperature, is it CRP count or something else?

Knowing what it is, and the scale of the measurement, would make it easier to teach.

From https://github.com/swcarpentry/lesson-template/blob/gh-pages/tools/check.py:

• Heading at line 56 should be level 2
• Document contains heading not specified in the template: Command-Line Arguments
• Document contains heading not specified in the template: Handling Multiple Files
• Document contains heading not specified in the template: Handling Command-Line Flags
• Document contains heading not specified in the template: Handling Standard Input
• Could not find the linked asset file ../../rules.html in /tmp/python/../../rules.html. If this is a URL, it must be prefixed with http(s):// or ftp://
• Could not find the linked asset file earlier in /tmp/python/earlier. If this is a URL, it must be prefixed with http(s):// or ftp://
• The topic page should not have sub-headings outside of special blocks. If a topic needs sub-headings, it should be broken into multiple topics.

Using the sys module for the command-line is a little bit messy. I think it would be better to use the more powerful, and scalable command-line parser.

https://docs.python.org/2.7/library/argparse.html

I could rewrite the lesson to use this library if people think this is a good idea.

There are some places where instructors and students using the Jupyter notebook need to do something that those in an interpreter don't need to do; e.g., %matplotlib inline. We should put these notes in callouts. It would help, for those not using the Jupyter notebook, to be able to visually ignore these callouts. We could do that by making the Jupyter callouts a specific style / color etc.

So we have a bunch of csv files with inflammation data. In 05-cond and 06-func, we develop some tests to tell whether or not our data is suspicious. One of the tests concerns the max behavior and one of the tests concerns the min behavior. But when we run detect_problems() on all of the files in inflammation*.csv we see that none of the csv files results in the "Seems OK!" print statement.

I just taught this lesson and some of the students stopped me to ask "Wait, so none of our data is okay?" I think it was demotivating for some of them (as it would be in the real world if you realized that all of your data were bad).

Should we consider changing either the datasets or the conditional tests in detect_problems() so that at least some of the inflammation data is considered OK?

The errors_01.py and errors_02.py modules for the lesson on "Errors and exceptions" are available in the code/ folder:

https://github.com/swcarpentry/python-novice-inflammation/blob/d79d0440540a30d7e839f82e370412e595e3549d/code/errors-01.py

We might point to them in the lesson so people can find them.

We should have at least one exercise at the end of each lesson.

Now that we have broken up our lessons do we plan on teaching in a single notebook or a notebook per lesson? If we are going to teach in a notebook per lesson there should be import statements at the beginning of each lesson for modules we use over and over again (such as Numpy).

@moorepants and @zonca reported that

We tried the 5.3 Beta Python and it was a little rough around the edges. (...) it has big issues, like functions being introduced in lesson 4 but used in 2.

From https://github.com/swcarpentry/lesson-template/blob/gh-pages/tools/check.py:

• Document contains heading not specified in the template: Syntax Errors
• Document contains heading not specified in the template: Variable Name Errors
• Document contains heading not specified in the template: Item Errors
• Document contains heading not specified in the template: File Errors
• The topic page should not have sub-headings outside of special blocks. If a topic needs sub-headings, it should be broken into multiple topics.

In "Analyzing Data from Multiple Files" python-novice-inflammation/04-files.md line 52 there's a long script for looping through files and showing three plots.

The script includes blank lines for readability. The blanks mess up the indentation when the script is run in ipython. The for loop prints three filenames, then executes subsequent lines once.

Please see my pull request.

In preparing for a workshop in which I will point to this material, I was working through the lessons at https://swcarpentry.github.io/python-novice-inflammation/01-numpy.html. I ran into problems, which I started to annotate using hypothes.is (e.g., https://via.hypothes.is/https://swcarpentry.github.io/python-novice-inflammation/01-numpy.html).

This morning, I then encountered http://www.software-carpentry.org/v5/novice/python/01-numpy.html, which seems to have been a differe (and perhaps more correct) version of https://swcarpentry.github.io/python-novice-inflammation/01-numpy.html.

Please make clearer the different versions of the material on the web.

We have a zip file in the main directory to make it easy to get access to the necessary files used in the lesson. There's repetition here, but not repetition that we can avoid for pragmatic reasons. But, we should make sure that this zip always contains the right versions of the right files, so it would be helpful to have a little Python or Bash script that either checks that this zip is correct, or perhaps ideally, creates the appropriate zip file (which should be identical to the existing zip file if the files in it are the same).

At line 39 of 01-numpy.md , we load our first data file:

numpy.loadtxt(fname='inflammation-01.csv', delimiter=',')

The .csv file is in the data directory. Since I was working from the top level directory, the file wasn't found. I had to change it to:

numpy.loadtxt(fname='data\inflammation-01.csv', delimiter=',')

Are students expected to work from the data directory?

See also the discussion on issue #4 .

There are no instructions for how to start Python. We jump right into import numpy in section 1.

I could copy the instructions from python-novice-turtles . Thoughts?

Fromhttps://github.com/swcarpentry/lesson-template/blob/gh-pages/tools/check.py:

• Document contains heading not specified in the template: For Loops
• Document contains heading not specified in the template: Lists
• Document contains heading not specified in the template: Processing Multiple Files
• The topic page should not have sub-headings outside of special blocks. If a topic needs sub-headings, it should be broken into multiple topics.

http://swcarpentry.github.io/python-novice-inflammation/06-func.html

detect_problems is defined using an underscore and called using camel case.
Propose to stick with PEP8 rules and make it an underscore.

This is also a bug in the lesson.

Testing is one of the hardest things for scientists to apply to their everyday coding. Our example of testing should make use of the scientific case we are following in this lesson to demonstrate how testing can be applied to scientific analysis.

It is unclear from the text of cmdline.md whether students are supposed to create files with the scripts, use the scripts in the code directory, or just read the code that's on the screen. This should be clarified.

I can not find a reference to this file anywhere else in the code base.

at the end of the lesson cmdline.md we modify our code accept stdin but then we only test our finished code on passing a filename. We should test stdin before we declare our code a success.

I've been working on adding solutions to the instructor's guide, but as I have never actually taught the Python+command line section, I don't have the solutions easily at-hand. It would save me some time to have a place to start rather than coding them all from scratch. Can anyone point me to a relevant notebook or repo that woudl help? Thanks.

Many of our callouts add additional information to the lesson (and are great for people reading through the lessons) but may not be necessary to teach during workshops. It could be useful to note that the callouts aren't necessarily things that need to be taught in a workshop, but are primarily there for interest or for further reading. A few ways we could do this (no real preference toward one; comments with preferences welcome!)

1. Add a note to the instructor guide specifying that callouts don't need to be taught unless it seems useful.
2. Add a callout on the index page explaining what callouts are used for (I think these appear at the start of many technical books like the "for Dummies" series).
3. Make all callout boxes collapsed by default, but shown when an arrow is clicked (as in, e.g., Bootsrap collapse)
4. Make a separate "optional" callout that is collapsed by default, but shown when an arrow is clicked.
5. Make a separate "optional" callout that is rendered differently than the usual blue callout.

The learning objectives for this section include "Explain the similarities and differences between tuples and lists." but this isn't really covered except briefly in the last exercise. Tuples are also mentioned later on (in 08-defensive.md).

Many of the python modules in code include hyphens in the name, making them practically impossible for learners to import (there is a way, but it's really obtuse and bad form). These should be changed to e.g. underscores in order to make them easily importable.

The first lesson for the python lesson [1] shows importing numpy and matplotlib.pyplot without an alias. but in the next lesson where we go back to plotting inflammation data [2]
we import glob but the code in the for loop use np and plt aliases.

If students follow the current lesson material, they would not know to import using an alias and will trip up when the code fails

[1] http://swcarpentry.github.io/python-novice-inflammation/01-numpy.html
[2] http://swcarpentry.github.io/python-novice-inflammation/04-files.html

From https://github.com/swcarpentry/lesson-template/blob/gh-pages/tools/check.py:

• Document contains heading not specified in the template: Image Grids
• Document contains heading not specified in the template: Conditionals
• Document contains heading not specified in the template: Nesting
• Document contains heading not specified in the template: Creating a Heat Map
• The topic page should not have sub-headings outside of special blocks. If a topic needs sub-headings, it should be broken into multiple topics.

I am preparing for teaching next week, and I have some questions:

1. Throughout the lesson, it is indicated that there is something wrong with the data. I have yet to figure out what exactly the problem is?
2. A more teaching technical thing: in 02 Checking our data, it is indicated that you can stuff the shown if statement into the for loop shown in 01 Processing multiple files. Do people do that when they teach, or do they just show them on screen and explain?
3. When I do put that if statement in there, I get either suspicious maxima or adding up to zero on all of them. Is this actually correct?

From https://github.com/swcarpentry/lesson-template/blob/gh-pages/tools/check.py:

• The document links to motivation.html, but could not find the expected markdown file /tmp/python/motivation.md
• The document links to discussion.html, but could not find the expected markdown file /tmp/python/discussion.md
• The document links to instructors.html, but could not find the expected markdown file /tmp/python/instructors.md

In 05-cond.md, this section:

Tuples and exchanges
Explain what the overall effect of this code is:

temp = left
left = right
right = temp

produces this error:

---------------------------------------------------------------------------
NameError                                 Traceback (most recent call last)
<ipython-input-199-42933dda7154> in <module>()
----> 1 temp = left
      2 left = right
      3 right = temp

NameError: name 'left' is not defined

left and right should be initialized to prevent the error.

Pull request here.

Hi all,

There's some hints here that @gvwilson wasn't pleased with the Python 3 transition, but that he now believes it is no harder to teach than Python 2. (I haven't found a primary source for these comments.)

I think every beginner we teach Python 2 puts additional load on the community (which has to maintain Python 2 code for that much longer), so my view is that we should port this lesson to Python 3 ASAP.

If there is interest, I'm more than happy to take this on.

Thanks!

The first lesson starts right off showing Python commands and there's never discussion of where someone following along should put those commands. This may be intentional so that the lessons are flexible, but it also leaves someone following along on their own confused as to where they should be typing. I'd recommend adding some preliminary discussion pointing people at the IPython notebook and terminal.

In

if 4 > 5:
    print 'A'
elif 4 =< 5:
    print 'B'
elif 4 < 5:
    print 'C'

The first elif should be ==

Pull request for fix is here.

https://twitter.com/hadleywickham/status/643381054758363136

There should be solutions provided to the exercises in the instructor notes so that instructors know the intended complete answer.