init

4 files changed, 186 insertions, 0 deletions

diff --git a/ara2influx.py b/ara2influx.py
new file mode 100644
index 0000000..26d7d0a
--- /dev/null
+++ b/ara2influx.py
@@ -0,0 +1,61 @@
+#!/usr/bin/env python
+""" usage: fetchday [options] DB
+
+Options:
+    --db FILE               path to db [Default: db.json]
+    --cantine CANTINE       name of the cantine [Default: thales-deutschland]
+    --host HOST             hostname or ip of influx [Default: omo]
+"""
+import sys
+from docopt import docopt
+import logging
+from influxdb import InfluxDBClient
+
+
+import json
+from datetime import datetime
+logging.basicConfig(level=logging.DEBUG)
+log = logging.getLogger()
+
+
+def db2points(db,cantine):
+    for day,things in db.items():
+        time = datetime.strptime(day,"%Y-%m-%d")
+        time = time.replace(hour=12)
+        for thing in things:
+            ret = {
+                "time": time.isoformat(),
+                "measurement":"dish",
+                "tags": {
+                    "allergies": ",".join( sorted(thing['allergies'].keys()) ),
+                    "additives": ",".join( sorted(thing['additives'].keys()) ),
+                    "title": thing["title"],
+                    "type": thing["type"],
+                    "description": thing["description"],
+                    "cantine": cantine
+                },
+                "fields": {
+                    "price": thing["price"]
+                }
+            }
+            log.debug(ret)
+            yield ret
+
+def main():
+    args = docopt(__doc__)
+    dbname = args['DB']
+    host = args['--host']
+    cantine = args['--cantine']
+    client = InfluxDBClient(
+            host=host,
+            port=8086,
+            database='araprice')
+    with open(dbname,'r+') as f:
+        db = json.load(f)
+        l = list(db2points(db,cantine))
+        # client.drop_database("araprice")
+        client.create_database("araprice")
+        client.write_points(l)
+
+if __name__ == '__main__':
+    main()
diff --git a/day.py b/day.py
new file mode 100755
index 0000000..4364762
--- /dev/null
+++ b/day.py
@@ -0,0 +1,104 @@
+#!/usr/bin/env python3
+""" usage: fetchday [options]
+
+Options:
+    --db FILE       path to db [Default: db.json]
+    --cantine NAME  name of the cantine (matches mein.aramark.de/NAME/) [Default: thales-deutschland]
+"""
+import sys
+from docopt import docopt
+import requests
+from bs4 import BeautifulSoup
+import logging
+
+import json
+from datetime import datetime
+logging.basicConfig(level=logging.DEBUG)
+log = logging.getLogger()
+from itertools import tee
+cantinedb = {
+        'thales-deutschland': 'aramark_menu2-2',
+        'buerocampus-wangen': 'aramark_menu2-3',
+        'ibm-boeblingen': 'aramark_menu2-3',
+        'metabo-nuertingen': 'aramark_menu2-3',
+        'vodafone-campus': 'aramark_menu2-2',
+        'continental-villingen': 'aramark_menu2-3'
+        }
+def get_data(cantine):
+    headers = {
+        'User-Agent': 'User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:58.0) Gecko/20100101 Firefox/58.0',
+        'Referer': f'http://mein.aramark.de/{cantine}/menuplan/',
+        'X-Requested-With': 'XMLHttpRequest',
+        'Cache-Control': 'max-age=0'
+    }
+    data = {
+        'action': 'aramark_menu2_content',
+        'id': cantinedb[cantine]
+    }
+    ret = requests.post(f'http://mein.aramark.de/{cantine}/wp-admin/admin-ajax.php',data=data)
+    if ret.status_code != 200:
+        log.error(ret)
+        raise ValueError('Return code is not 200')
+    return ret.content.decode()
+
+def get_data_test():
+    return open("./test.html").read()
+
+def get_things(row):
+    for item in row.select('.aramark_menu2-counter'):
+        try:
+            thing = {
+                "allergies": {},
+                "additives": {},
+                "type" : item.select_one('.aramark_menu2-countertitletop').find(text=True),
+                "description" : item.select_one('.aramark_menu2-menutext > p').find(text=True),
+
+            }
+            try:
+                price = float(item.select_one('.aramark_menu2-menuprice > .aramark_menu2-menuprice').find(text=True).replace(',','.').replace('EUR','').strip())
+            except:
+                price = float(thing['description'])
+            thing['price'] = price
+            try:
+                thing['title'] = item.select_one('.aramark_menu2-countermenu h4').find(text=True)
+            except:
+                thing['title'] = thing['type']
+
+            entries = item.select('.aramark_menu2-menuinfoallergenes  li span')
+            for i in range(0,len(entries),2):
+                thing['allergies'] = { entries[i].find(text=True): entries[i+1].find(text=True) }
+
+            additives = { }
+            entries = item.select('.aramark_menu2-menuinfoadditives li span')
+            for i in range(0,len(entries),2):
+                thing['additives'] = { entries[i].find(text=True): entries[i+1].find(text=True) }
+            yield thing
+        except Exception as e:
+            log.error(f"problem with {e}")
+            log.debug(item.prettify())
+def main():
+    args = docopt(__doc__)
+    dbname = args['--db']
+    cantine = args['--cantine']
+    try:
+        with open(dbname,'r+') as f:
+            db = json.load(f)
+    except:
+        log.info('creating new db')
+        db = {}
+
+    updated = datetime.now().isoformat().split('T')[0]
+    soup = BeautifulSoup(get_data(cantine), 'html.parser')
+    # soup = BeautifulSoup(get_data_test(), 'html.parser')
+    for row in soup.select('.aramark_menu2-menus > div'):
+        day = row.get('data-aramark_menu2-day')
+        things = list(get_things(row))
+        db[day] = things
+
+    log.info(f'writing db {dbname}')
+    with sys.stdout if dbname == '-' else open(dbname,'w+') as f:
+        json.dump(db,f,indent=2,sort_keys=True)
+
+
+if __name__ == '__main__':
+    main()
diff --git a/day.sh b/day.sh
new file mode 100644
index 0000000..6780dd9
--- /dev/null
+++ b/day.sh
@@ -0,0 +1,11 @@
+#!/bin/sh
+OUTDIR=${OUTDIR:-$PWD/db}
+# old: thyssenkruppsystemengineering
+# empty: metabo-nuertingen
+mkdir -p $OUTDIR
+for i in thales-deutschland ibm-boeblingen buerocampus-wangen continental-villingen vodafone-campus ;do
+  db=$OUTDIR/$i.json
+  python3 day.py --db $db --cantine $i
+  python3 ara2influx.py $db --cantine $i
+  # git add $db
+done
diff --git a/week.sh b/week.sh
new file mode 100644
index 0000000..f972e16
--- /dev/null
+++ b/week.sh
@@ -0,0 +1,10 @@
+NOW=$(date -I)
+OUTDIR=${OUTDIR:-$PWD/pdf}
+mkdir -p $OUTDIR
+# Broken:
+# vodafone-campus thyssenkruppsystemengineering
+# Gone? metabo-nuertingen 
+for i in thales-deutschland ibm-boeblingen buerocampus-wangen continental-villingen ;do
+  db=$OUTDIR/$i-week.pdf
+  wget http://gast.aramark.de/$i/menu/pdf/woche_de.php -O $db
+done