Compare commits

..

No commits in common. "ee27d7b679f9179d93c7d12cc51e1cb7876ed7a4" and "1e7ac17d903a136674146a0f9a2ea94cca93d2d4" have entirely different histories.

4 changed files with 6 additions and 21 deletions

View File

@ -1,16 +1,16 @@
FROM debian:11 FROM python:3.8
RUN apt-get update && apt-get install -y git curl libcurl4-openssl-dev build-essential vim libssl-dev python2 python-dev wget python3 python3-pip RUN apt-get update && apt-get install -y git curl libcurl4-openssl-dev build-essential vim libssl-dev python-pip
RUN wget https://bootstrap.pypa.io/pip/2.7/get-pip.py && python2 get-pip.py
RUN python2 -m pip install cqlsh RUN pip2 install cqlsh
RUN addgroup appgroup && \ RUN addgroup appgroup && \
adduser appuser && adduser appuser appgroup adduser appuser && adduser appuser appgroup
RUN mkdir /app /src RUN mkdir /app /src
ADD requirements.txt /src/ ADD requirements.txt /src/
RUN CASS_DRIVER_BUILD_CONCURRENCY=4 pip3 install -r /src/requirements.txt RUN CASS_DRIVER_BUILD_CONCURRENCY=4 pip install -r /src/requirements.txt
WORKDIR /src WORKDIR /src
ADD . /src ADD . /src
RUN pip3 install /src/ RUN python /src/setup.py install
WORKDIR /app WORKDIR /app

View File

@ -1 +0,0 @@
docker build . -t dr.kemt.fei.tuke.sk/websucker:dev

View File

@ -158,13 +158,6 @@ def check(ctx,domain):
res = db.check_domain(domain) res = db.check_domain(domain)
print(res) print(res)
@cli.command(help="Export domain as JSON doc per line")
@click.pass_context
@click.argument("domain")
def tojson(ctx,domain):
db = create_database_from_context(ctx)
db.export_domain(domain)
@cli.command(help="Print daily report") @cli.command(help="Print daily report")
@click.pass_context @click.pass_context
def report(ctx): def report(ctx):

View File

@ -1,6 +1,5 @@
import cassandra import cassandra
import cassandra.cluster import cassandra.cluster
import cassandra.query
from cassandra.auth import PlainTextAuthProvider from cassandra.auth import PlainTextAuthProvider
import random import random
import os import os
@ -9,7 +8,6 @@ import datetime
from websucker.parser import normalize_link,urlunparse from websucker.parser import normalize_link,urlunparse
import collections import collections
import math import math
import json
VERSION = "sucker6" VERSION = "sucker6"
@ -545,11 +543,6 @@ INSERT INTO content(
result.append((d,0)) result.append((d,0))
return result return result
def export_domain(self,domain):
rows = self.session.execute("SELECT JSON * from content WHERE domain_name=%s",(domain,))
for row in rows:
print(row[0])
def get_visit_links(self,domain,recent_count,old_count,random_count): def get_visit_links(self,domain,recent_count,old_count,random_count):
dblinks = [] dblinks = []
rows = self.session.execute("SELECT url_schema,url_path,url_query,update_time FROM links Where domain_name=%s AND link_status='seen'",(domain,)) rows = self.session.execute("SELECT url_schema,url_path,url_query,update_time FROM links Where domain_name=%s AND link_status='seen'",(domain,))