Spacelog

This is the source code for Spacelog, a website for experiencing space missions through radio transcripts and photography.

With the exception of the font and some icons (credited on the mission about page), everything outside the missions directory is released under the CC-0 license. Mission images are credited in the mission's _meta file.

We hope you have fun with this -- we have!

The Spacelog team

Getting involved without technical knowledge

Correcting minor errors

For small errors (whether transcription errors, or something like spelling mistakes on the rest of the site), it's probably easiest to just email them through to us at spacelog@googlegroups.com.

Helping transcribe a new mission

If you download the PDFs from NASA for a mission you want to add, you'll discover that you can select text in them and copy it out into a text editor, or something like OpenOffice Writer, Apple's Pages or Microsoft Word. You'll see lots of lines that look like this:

02 07 55 20 CMP I believe we've had a problem here.

However some of the lines will have errors (from the small, such as O instead of 0, to the large such as entire lines being completely garbled). If you go back to the PDF, you can usually quite easily figure out what was originally typed out but which the automatic OCR didn't get right.

There are also some non-dialogue lines. These should all be indented by a single tab; the most important ones are:

    TAPE 2/1
    PAGE 9

which happen at the start of a new page of the PDF. In this case they mean that it's the first page of the transcript of tape 2, and is page 9 of the complete transcript. We particularly need the page number so we can link back to the original typescript in the site.

If you clean up the text version like this, and send it through to us, we can do the rest (although it may take us a bit of time). Since some of the missions are quite long (hundreds of pages of transcript), you may want to share the load between a group. Whether you do that or decide to go it alone, it'd be great if you could let us know at at spacelog@googlegroups.com what you're working on, so we can help you out, and make sure you don't duplicate others' effort.

A quick note about multiple transcripts

For Mercury-Atlas 6, there is only one transcript available, that of the air-to-ground radio communications (John Glenn's mike was hot through the entire mission). For Apollo 13, there is a second transcript from the Command Module recording, but it cuts out very early into the mission, so we didn't consider it worth including.

However for many other missions there are multiple transcripts. If you're adding missions to Spacelog, please keep these transcripts in different files. We don't yet have support to identify them distinctly, but if you move them all into one big file it'll be impossible for us ever to work on that!

Getting involved with technical knowledge

Getting set up

Source code

Clone the repository from git:

$ git clone git://github.com/Spacelog/Spacelog.git

However for any changes you make (fixed, new missions, or even new website features), you will want to issue a pull request to us from another github repository). In order to do that, you'll need to set up a github account, and while logged in go to our repository there and hit the "fork" button (top right, near the search box). This will create a copy of Spacelog under your github user; you can then grab the SSH URL (which will look like git@github.com:<your user>/Spacelog.git) and use for git clone, as:

$ git clone <github SSH URL>

You can then make changes, commit them to your local copy (git commit), push them up to your github copy (git push) and finally send us a pull request (which you do via the github website). Github has some great guides to getting started with git and github linked from their homepage once you're signed in, in particular their description of forking a repository.

Software to install

python (and pip)
redis (often packages as python-redis; we need Redis 2.0 or later)
Xapian and its python bindings (the search engine library we use; often packaged as python-xapian)
Sass (a Ruby preprocessor for CSS; gem install sass)
various python modules (run pip install -r requirements.txt)
imagemagick and optipng (for building the stats images on the phase pages; this is optional)

The easiest way to grab the python modules may be to build a virtualenv in the Spacelog checkout:

virtualenv --system-site-packages ENV

pip install --upgrade pip ENV/bin/pip install -r requirements.txt

then use ENV/bin/python where normally you'd use just python. (If it's recent enough, virtualenv may install a script to "activate" its copy of python in your shell, which you run as source ENV/bin/activate, or ENV/Scripts/activate.bat on Windows.)

Running the code

If you have screen installed and are using a virtualenv as above, you should be able to just run make screen to get everything running for you. You also need to run make reindex to load all the details of the missions. Then you can point your web browser at http://dev.spacelog.org:8001/ and the global homepage should come up; from there you can navigate to other missions, which will appear at URLs such as http://apollo11.dev.spacelog.org:8000/. The DNS is managed by us, and providing you're online everything will just work.

make screen fires up an instance of screen, which is an easy way of running multiple programs on one terminal. Currently it leaves you looking at the development server log for the global homepage, but you can switch to a blank terminal by typing ^A 0, ie: holding down the ctrl key, pressing A, releasing ctrl and then pressing 0. This is a good place to run make reindex from.

All our make commands will take care of the virtualenv for you; if you're not using one, you can use PYTHON=python make <whatever> instead.

The details

If you can't use make screen, or simply if you wish to know how it all fits together under the skin, then here's the details. It's also helpful in case you're developing the code directly, since under certain circumstances the Django development server can crash, and will need restarting. Similarly if you add a new CSS file, you will currently have to restart the appropriate devcss server.

We use redis for storage, so you need to have redis-server running before you run make reindex in the checkout directory, which will import all the mission data into redis. You may also want to do make statsporn to build the graphs for the phases page of how much was said at different times (and, in case we've added more graphs but haven't updated this, other things :-).

You then need to have some other servers running on top of redis:

make devcss will run sass in watch mode, so changes to CSS files will be reflected automatically
make devcss_global will run sass for the project homepage
make devserver will run the mission-specific websites; if not using a virtualenv, PYTHON=python make devserver should do the trick
make devserver_global will run the project homepage; if not using a virtualenv, PYTHON=python make devserver_global should do the trick

Hosts setup for offline use

If you're not online, you can't use our development DNS, so you'll need edit /etc/hosts to include an alias dev.spacelog.org, plus aliases of the form <mission>.dev.spacelog.org, such as apollo13.dev.spacelog.org and mercury6.dev.spacelog.org; these all need to point to localhost (or to your virtual machine, if that's how you develop things). For instance, here's an /etc/hosts entry using localhost (put this in addition to the localhost line already in there):

127.0.0.1		apollo13.dev.spacelog.org mercury6.dev.spacelog.org dev.spacelog.org

and here's one for a virtual machine (you'll need to change the dotted quad at the start of the line):

192.168.56.101	apollo13.dev.spacelog.org mercury6.dev.spacelog.org dev.spacelog.org

Reindexing

Whenever you edit information about a mission, or add a new one, you need to run make reindex again. If you get errors you may find the lognag.pl script in mcshred/src useful: just give it some transcript files and it'll tell you where it finds possible errors or weirdnesses. (For new missions, you'll probably have to add things into the valid speakers list at line 71.)

Note that a full make reindex can take a while, so you can index just a single mission by doing ENV/bin/python -m backend.indexer ma6 or similar (or just python -m backend.indexer ma6 if you aren't using a virtualenv.

External Source Images

We make use of external source images (which we haven't created ourselves) in the form of:

.pngs of transcript PDF pages
Original NASA photographs

For reasons of size these aren't stored in git, they're stored in the spacelog Amazon S3 bucket (served by Cloudfront on http://media.spacelog.org). By default, our settings point you to this host. If you want to test adding your own images, you can change the MISSIONS_IMAGE_URL in website/configs/settings.py to serve them locally. File a github ticket if you need images uploaded to S3.

Adding a new mission

You'll need to create a directory in missions. For Mercury-Redstone missions these should start mr, for Mercury-Atlas ma, for Gemini they start just g and for Apollo a. If anyone wants to do non-NASA missions, or Shuttle missions, then get in touch and we'll figure out a naming convention.

Look in transcript-file-format for a description of how we lay out files. If you're transcribing a mission we don't have, you will find the example _meta and TEC files useful, since they are the main two files you'll need to create (if you're going to include more than just the air-to-ground transcription, you'll want to put that in TEC, the command module transcript in CM, and so on). If you can make them in that format (or get as close as you can), and send them through to us along with a link to the original transcript PDFs you used, we'll do the rest. (If you are gifted in design, the source files for all the artwork we've created is available, although we haven't yet put it online -- yell if you need it as a basis for making things like orbital diagrams.)

Multiple transcripts

As noted above in the information for non-technical folk, if you clean up multiple different transcripts for a single mission (for instance you might do not only the TEC ("technical" ground-to-air) recording but also the CM and/or LM recordings), then please keep them in separate files rather than merging them.

Technical glossary

Within the system, there are a number of terms that describe pieces of the system but do not necessarily match what is shown on the websites.

TRANSCRIPT FILE -- our textual representation of the original transcript; see transcript-file-format/TEC for a commented example
TIMESTAMP -- four colon-separated numbers that represent the GET (Ground Elapsed Time), the time since launch within the mission; the four numbers are days, hours, minutes, seconds, so ignition is 00:00:00:00; these are used in the transcript files, and also in URLs
LOG LINE -- smallest linkable chunk, identified by timestamp and transcript file
RANGE -- a range between two timestamps (can be the same two)
LABEL -- a keyword applied to a range within a specific stream (note that labels are not currently used)
META FILE -- a per-mission file called _meta that contains information such as glossary items, pull quotes for the homepage, and acts (see transcript-file-format/_meta for a commented example
CHARACTER -- a speaker who appears in a transcript file; additional information about them appears in the meta file
SHIFT -- a range where one "role" character (such as CAPCOM or the flight director) can be identified with a "real" character (such as Charlie Duke or Deke Slayton); ranges are defined in the meta file

From this we generate a number of higher-level pieces which are used in the website.

ACT -- an editorially defined range that represents a segment of the mission, which may for instance reference orbital mechanics (in the websites these are referred to as phases)
KEY SCENE -- an editorially defined point in the transcript where an important event or exchange starts
STREAM -- a collection of related content arranged on a timeline

Characters

Characters are defined in a _meta key characters, which is a dictionary with keys the character identifiers in the transcript and values a further dictionary of information about that character. For instance:

{
    "characters": {
        "P": {
            "role": "astronaut",
            "name": "Virgil Ivan Grissom",
            "short_name": "Gus Grissom",
            "mission_position": "Pilot",
            "bio": "A few sentence biography",
            "photo": "grissom.jpg",
            "photo_width": 190,
            "photo_height": 205,
            "avatar": "grissom.jpg"
        }
    }
}

This defines the character P. bio, photo (stored in the mission's images/people directory; photo_width / photo_height should be set appropriately) are used on the people page.

role is based on initial usage, and so can be a little confusing. It should be one of astronaut, mission-ops, mission-ops-title or other (defaulting to other). Astronaut means a full-size, prominent place on the main people page (190x205 image with biography as above, and also support for stats and a quote); mission-ops-title will get a less prominent position on the main people page (190x205 with biography); mission-ops go on a second page (linked as "View Mission Control Team" from the main people page), where they get a 190x155 photo and brief biography.

The people pages show the full name (the name key) and the mission position from the character definition. The short name is shown within the transcript, with the avatar (48x48, stored in the mission's images/avatars directory; astronauts get a yellow hue to differentiate them from those not in space during the mission) alongside.

Character stats and quotes

Characters with a role of "astronaut" can optionally have statistics and quotes, as shown below:

{
    "characters": {
        "CDR": {
          "role": "astronaut",
          "name": "James A. (Jim) Lovell, Jr.",
          "short_name": "Jim Lovell (CDR)",
          "mission_position": "Commander",
          "bio": "...",
          "photo": "lovell.png",
          "photo_width": 190,
          "photo_height": 205,
          "avatar": "jim_lovell.jpg",
          "stats": [
            {
                "value": 715,
                "text": "hours in space"
            },
            {
                "value": 4,
                "text": "missions"
            },
            {
                "value": 42,
                "text": "age at launch"
            }
          ],
          "quotable_log_line_id": "TEC:05:18:04:46"
        }
    }
}

The quote must be in the transcript, and is given as the transcript name followed by the GET of the logline. (This means you can't use loglines that have multiple speakers.)

There should be three stats, and you will likely have to juggle things around in order to make them fit the layout. We haven't used stats on all missions; it isn't always possible to find suitable figures for the astronauts involved.

The shift system

On longer missions, generic positions such as CAPCOM or F (flight director) are shared between several people operating in shifts. This is done by having a character dictionary key of shifts, whose value is a list of two element lists:

{
    "characters": {
        "STONY": {
          "role": "other",
          "name": "Blockhouse Comm",
          "short_name": "Stony",
          "shifts": [
            [ "DEKE_SLAYTON", "00:00:00:00" ]
          ]
        }
    }
}

This means that the first shift is taken by the character with identifier DEKE_SLAYTON, at GET 00:00:00:00. Since identifying shifts at this remove from the event isn't always straightforward, there will often be a third element in the list giving an annotation, justification or source:

{
    "characters": {
        "CC": {
            "role": "mission-ops-title",
            "name": "Capsule Communicator",
            "short_name": "CapCom",
            "bio": "...",
            "photo": "capcom.jpg",
            "photo_width": 190,
            "photo_height": 205,
            "avatar": "capcom_generic.png",
            "shifts": [
              ["JOE_KERWIN", "-00:01:00:00", "strictly, only Kerwin, Brand and Lousma were taking shifts (AFAICT), however other astronauts come on as CAPCOM in the original transcript, and we use the shift mechanism to display that properly"],
              ["JOHN_YOUNG", "00:04:39:01", "identified by PAO transcript"],
              ["JOE_KERWIN", "00:04:50:45"],
              ["VANCE_BRAND", "00:07:09:09"],
              ["JACK_LOUSMA", "00:16:00:00", "uncertain (and moot) since he doesn't appear in the transcript at this point"]
            ]
        }
    }
}

We also (as in the first example above, from Gus Grissom's Mercury-Redstone 4 flight) use the shift system to "delegate" a generic character (such as STONY, the callsign for an astronaut communicator in the blockhouse during Mercury launches) to a specific character (in this case Deke Slayton) who served in that role for the mission in question.

Code layout

The main code is two Django projects and a python library for managing transcript files into a redis data store. There is also a directory full of per-mission information (transcript files, images and so on), and some other tools directories.

website/ runs the per-mission websites (Django project)
global/ runs the project global homepage (Django project)
backend/ (python library to load transcript files into redis/xappy, generate stats images, and provide an API for accessing streams and other information)
transcript-file-format/ (documentation of the transcript file format)
missions/ contains the per-mission data, particularly the transcript files and meta file, but also images and so forth
tools/ (standalone python tools)
mcshred/ (python and perl programs for dealing with OCR data from NASA PDFs)
ext/ (historical mechanism used during development because pip doesn't work in forts)

jweather/Spacelog