About Venera

venera - is a project for search interesting girls in various dating platforms. It's developed for automate swiping and analyzing profiles.

Features

• Web control interface;
• Can run multiple search tasks with different accounts/search settings/dating platforms;
• Implements Telegram Bot for control and notifications;
• Text analyzer for profile's description with patterns;
• Machine Learning for rate girls photos;
• A lot of tricks for avoiding ban;
• Can work as backend daemon;
• Ability of download database with photos (ex. for post-analyze or use as dataset for train neural network).

Supported dating platforms:

• Tinder (supports auth only via Google Account);
• Mamba (needs to patch fresh App ID);
• Badoo DEPRECATED.

Dependencies

• PostgreSQL (used as database);
• libtensorflow 1.14.0;
• google-chrome;
• libx11-xcb1, libxt6, libdbus-glib-1-2 (for Headless browser).

Build

git clone https://github.com/racoon-devel/venera
cd venera && mkdir tensorflow
wget "http://storage.googleapis.com/tensorflow/libtensorflow/libtensorflow-cpu-linux-x86_64-1.14.0.tar.gz" && \
   tar -zxv -C `pwd`/tensorflow -f libtensorflow-cpu-linux-x86_64-1.14.0.tar.gz && \
   rm -f libtensorflow-cpu-linux-x86_64-1.14.0.tar.gz

export LD_LIBRARY_PATH=`pwd`/tensorflow/lib
export LIBRARY_PATH=`pwd`/tensorflow/lib
go build app/venera/venera.go

Run

./venera -config <path to config> [-verbose]

Configuration

# Sample configuration file for venera

[http]
# endpoint IP address
ip = "0.0.0.0"
# endpoint port
port = 80
# HTTP request timeout
timeout = 5
# Admin login
username = "admin"
# Admin password
password = "admin"

[database]
host = "localhost"
port = 5432
user = "venera"
password = ""
database = "venera"

[directories]
content = "/var/lib/venera"
downloads = "/var/lib/venera/downloads"

[telegram]
# Bot API token
token = ""
trustedUser = ""

# Only SOCKS5 proxy
[proxy]
enabled = false
ip = "127.0.0.1"
port = 8080
user = ""
password = ""

• http - section of HTTP server settings;
• database - credentials to access a database;
• telegram - settings of Telegram Bot:
  • telegram.token - token of Telegram Bot
  • telegram.trustedUser - your Telegram user name.
• directories - paths to service directories;
  • directories.content - path to content directory with templates, pages, configs etc.
  • directories.downloads - path to directory which will be used as downloads dir for export database.

Steps to run

1. Set directories.content to absolute path of content directory from this repo;
2. Set directories.downloads to existing directory;
3. Create Telegram Bot Token and set telegram.token;
4. Set telegram.trustedUser to your own Telegram username;
5. Run Venera

./venera -config <path to config> [-verbose]

6. Go to "http://127.0.0.1:80/". Default credentials admin:admin.

Docker

1. Create Telegram Bot Token and set telegram.token in configs/venera.docker.conf;
2. Run

docker-compose up

How It works ?

For each dating platform you need to set some search settings. Usually there are login, password, location and keywords. Keyword means like or dislike. It supports text patterns. Examples:

смотр* аниме, игр* на гитар*@5, вело* 

игр* на гитар*@5 will be match with играла на гитаре, играю на гитаре etc. @5 - weight for keyword matching.

Rater - is an engine for rate user profile. It can be:

• default - rate profile text description only (via keywords);
• ml - rate only photos;
• default+ml - both.

Rater grade each fetched profile and put rating and all information do database. Tinder implementation provide Telegram notification about top profiles every day.

Tunning

For each dating platform you can tune some weights or factors. Ex. Tinder config (content/configuration/default.tinder.conf):

MinBioLength = 50
RelevantDays = 25

VIPAccountWeight = 2
BioMatchWeight  = 2
BioPresentWeight = 5

AlcoFactor = 1
SmokeFactor = 1
BodyFactor = 1

LikeThreshold = 40

• MinBioLength - minimal length of profile text description. It means if profile description is too short, profile will be passed and ignored;
• VIPAccountWeight - add or remove scores if profile have Premium or Plus;
• BioMatchWeight - weight of matched keyword;
• BioPresentWeight - add scores if user profile text description filled;
• Alco/Smoke Factor - multiplier for bad habits( <0 - negative, 0 - ignore, >0 - positive);
• LikeThreshold - if person have grade >LikeThreshold, she will receive a like.

For retrain neural network you should do this and put result to content/ml.