This is for Ingress. If you don't know what that is, you're lost.
This is a heavily modified original maxfield software that will generate an easy-to-follow fielding plan using Google Spreadsheets. The benefits over the original maxfield program are:
The main perk of this implementation is the Google Spreadsheet fielding plan, which provides easy-to-follow step-by-step instructions for the agent. This is how it looks on a mobile phone:
Here are a few examples of annotated spreadsheets generated with fieldplan:
Open those links on your phone and choose "Use the App" -- then switch to the plan tab and zoom in for best readability.
Fieldplan does not touch any of the Niantic's servers, so it is perfectly within the Terms of Service. All of the data comes from the spreadsheet and from the Google Directions API.
The main reason why I hacked on maxfield is to make it more convenient for biking, as having a simple plan to follow allows me to concentrate more on biking and less on figuring out what to do next.
This is a console python application. It expects a POSIX-compatible system (Linux, OS X) and a virtualenv-3 setup. After initializing the environment, you can install the required libraries using:
pip install -r requirements.txt
If you've never used a console, python and pip before, then you'll have a bit of a hard time at first, but it's not that hard to learn.
To start generating fieldplans, you will need to first get a credentials.json file and then generate a token. It's a bit annoying and complicated, but you only have to do this once.
This is also annoying and complicated, and you also have to only do this once.
CAUTION: Google will require you to set up billing for your project, so if you're not in a position to put in a credit card, then you shouldn't bother with this. You should not get charged unless you're making many thousands of API calls daily. Using your Directions API key with ingress-fieldplan should be effectively free for you if you're not calculating hundreds of plans every hour. Fieldplan also relies heavily on caching, so if we've looked up the walking/biking/driving distance between two portals once, it will be stored in local cache for all future lookups.
If you do set up your Google Directions API key, then you will greatly benefit from much more accurate distances, especially for portals that are in close proximity but require long detours.
-g {yourkey}
switch onceOnce you have your credentials.json (and your Directions API key, if you choose), you can run the following command to test out if it's working:
./fieldplan -n -s https://docs.google.com/spreadsheets/d/1TbwOCNpsvA7CjOTPv_98Iirjt_siOoAgoTKa0PTglgU/edit
If it didn't crash horribly, then you're in business!
Once you have all the portals entered, copy the spreadsheet URL and run the command:
./fieldplan -s https://docs.google.com/spreadsheets/d/xxx/edit#gid=0
Depending on how many portals you have in the spreadsheet, it will take anywhere from a few seconds to 10-15 minutes to run with the default set of iterations. The results will be saved as a new sheet and cached on your computer, so if you run the same spreadsheet again, it will continue from the previous best plan.
Look at the output of
./fieldplan --help
to find all the knobs and levers you can tweak. Here are a few pointers:
The default is 5,000 random iterations to find the best fielding plan. The bisecting and fielding is done randomly, largely because finding efficient movement plans for a set of geographical coordinates is one of those "NP hard" problems (look up the "Travelling Salesman Problem"). There is an optimization step after each random plan to fix the worst inefficiencies, so the results after each iteration are already tweaked. In my personal experience using it, I've found 10,000 a good number of iterations to generate decent plans.
Generally:
Since iterations are largely random, it's entirely possible to find the best possible plan on your first run, and to only find terrible plans even after 10,000 iterations. YMMV.
Getting lots of keys used to be difficult, but really isn't any longer. If you're good at glyphing, then you can expect to get 2 keys almost each time you hack a portal.
Your strategy, therefore, should be to always glyph-hack with "More". If you're capturing by yourself, that should limit you to 3-glyph portals, and that's not too hard to do -- you can even throw in a "Complex" to speed things up.
Because of the optimization routines, you may end up with plans that require you to get lots and lots of keys from a single portal. This may seem crazy, but if you are using Heat Sinks, this actually results in faster gameplay than plans where you need 3-4 keys from every portal.
For example, if you have a portal requiring 8 keys, you would:
As you see, getting as many as 8 keys usually requires a single RHS mod and about 6 minutes of waiting for portal cooldown. This ends up much faster than installing common Heat Sinks at every other portal to get 3-4 keys, and you end up using fewer Heat Sink mods.
Note, that the plan instructions will always tell you how many keys you need for the portal before you leave. Often, even if a portal requires 5-6 keys, you may not need to get them all at once.
While estimating the time to play, the software will assume that you will get
1.5 keys per hack and use Rare Heat Sinks to speed up portal cooldown.
If you only have regular Heat Sinks, you can specify that with --cooling hs
.
Other options are:
rhs
: Rare Heat Sink (default)hs
: Heat Sinkvrhs
: Very Rare Heat Sinknone
: don't use Heat Sinks at allidkfa
: you have all the keys and hack/cooldown times should not be countedRunning with --cooling none
is recommended if you have lots of time, don't
mind extra moving around, or don't want to spend your Heat Sink mods.
You will probably be planning your field ops either from home, on the way from home to work/school, or from a parking/transit stop location. To generate plans that are more efficient with those locations, you should add them as waypoints to your spreadsheet.
Use special name indicators at the start of the portal names:
#!s Location Name
for your start waypoint#!e Location Name
for your end waypointFor example:
#!s My Home
https://intel.ingress.com/intel?ll=45.498803,-73.598872&z=21
#!e My School
https://intel.ingress.com/intel?ll=45.504427,-73.574309&z=21
Waypoints can be either at the start or at the end of the portal list.
You can also add portals you need to visit to destroy blockers by using the
same logic as with start/end waypoints. Use the #!b Portal Name
indicator
in the left column to mark that a portal is a blocker and not part of the
fielding plan.
Note: The software has no idea where the blocking links are, so you will need to review the plan to make sure that you are not throwing early links before destroying the blockers that would be in the way.
By default, fieldplan will try to maximize AP per minute of gameplay, but
using the -u
switch you can tell it to consider field sizes as well, in an
attempt to find plans that would also give you highest area capture per
minute of gameplay.
Note: the software has no way of knowing the actual in-game MU density, so it will simply give higher priority to larger fields.
Passing the -p
switch will generate a set of step-by-step PNG files that
allows you to preview the plan in action. Here's what it is for the Biking
example above:
You may need to install python-tkinter for it to work.
Note: This is an experimental feature and currently requires significantly
more iterations to find efficient plans, so run it with -i 50000
and higher.
This probably happened to you -- you found an area with lots of uncaptured
portals, but you only have a limited amount of time to play. You can give
Fieldplan that large list of portals and ask it to find you a plan that would
give you maximum AP (or MU, with -u
) within the time constraint specified.
The software will try various subsets of portals until it finds something that
satisfies the parameters.
For example, there's a historical site with 25 portals, but fielding them all would take over 3 hours:
--maxtime 120
Fieldplan will try to find the most efficient plan that will take no more than 2 hours to execute.
Since Fieldplan prioritizes plans with highest AP (or MU) per minute of gameplay, it's possible that the most efficient plan it finds will contain only a few portals from the list. This especially tends to happen when prioritizing MU over AP.
You can pass --minap
to tell Fieldplan to not consider plans resulting in
too few total AP points. For example, to get plans with at least 50,000 AP, run
--minap 50000
.
Manually inputting portals can be tedious, so there is a way to copy and paste the list from IITC. You will need:
Here's how to use it:
XXX
in the "Polygon/TSV" columnA1
cellFieldplan needs Portal names in the column A
and Intel URLs in the column B
,
so you will need to either:
D
, C
, A
, orFieldplan will ignore anything not in columns A
and B
.
CAUTION: IITC is not an official resource provided by Niantic, and your use of it may be against their Terms of Service.
You can open a GitHub issue if something is not working for you, but please keep in mind that this is entirely a hobby project and I may not have a chance to help you out.
Happy fielding!
ENL agent: mricon