Store a copy of your last.fm listening history in a local sqlite database for safe keeping and offline analysis

Requires golang with module support (1.11+)

go build -o localfm cmd/main.go

Put the compiled binary wherever you like. Then create a new database:

sqlite3 $FOO.db < sqlite_schema.sql

Most configuration is done through environment vars.

DSN is similar to the format of sqlalchemy and other python tools. Since localfm only supports sqlite, it only ever needs to refer to a path:

export DSN="sqlite:https://$FOO.db"

LastFM API keys are required to retrieve updates:

export LASTFM_USERNAME="$YOURNAME"
export LASTFM_API_KEY="XXXXX"
export LASTFM_API_SECRET="YYYYY"

The web UI requires a little configuration also:

export LOGIN_USER="$YOURNAME"
export LOGIN_PASSWD="$BCRYPT_HASH"
export SESSION_SECRET="$RANDOM_32_BYTES"

The bcrypt password can be generated from a *nix shell:

htpasswd -bnBC 10 "" password | tr -d ':\n'

A session secret can be generated with:

python -c 'import random, string; print("".join([random.choice(string.ascii_letters + string.digits) for x in range(32)]))'

(works in either python 2 or 3)

last.fm streams occasionally contain repeated entries a few seconds apart which can affect your metrics. localfm can be configured to flag these repeated entries and exclude them from it's statistics. To enable this feature, set a threshold (in seconds) in the environment var LOCALFM_DUPLICATE_THRESHOLD. Any tracks that are repeated in this interval or less will be flagged.

export LOCALFM_DUPLICATE_THRESHOLD="5"

A sample.sh script is provided which can be customized.

Run localfm on a newly created database and it will download your entire listening history. Subsequent runs will do incremental updates of new activity since the last run. If there's an error, a checkpoint.json file should be written that allows the process to resume.