first stab at session-based fetching

This commit is contained in:
Rudis Muiznieks 2025-04-02 17:09:48 -05:00
parent f33c388e1f
commit 28e3eb2dc2
Signed by: rudism
GPG key ID: CABF2F86EF7884F9
2 changed files with 86 additions and 41 deletions

View file

@ -11,10 +11,12 @@
# limitations under the License.
from ovos_workshop.skills.fallback import FallbackSkill
import sseclient
import requests
import json
import html
import re
from urllib.parse import quote_plus
class KagiSkill(FallbackSkill):
@ -22,12 +24,50 @@ class KagiSkill(FallbackSkill):
self.register_fallback(self.handle_fallback_kagi, 80)
def handle_fallback_kagi(self, message):
if "kagiApiKey" not in self.settings:
if "kagiSession" in self.settings:
return self.get_using_session(message)
elif "kagiApiKey" in self.settings:
return self.get_using_api(message)
else:
self.log.error(
"kagi not configured yet, please set your kagiApiKey in %s",
"kagi not configured yet, please set kagiSession or kagiApiKey in %s",
self.settings.path,
)
return False # not configured yet
def get_using_session(self, message):
session_cookie = self.settings.get("kagiSession", "")
header_content = {
'Accept': 'text/event-stream',
'cookie': f'kagi_session={session_cookie}'
}
read_url = f'https://kagi.com/stream_fastgpt?query={quote_plus(message.data.get('utterance'))}'
try:
client = sseclient.SSEClient(read_url,
headers=header_content,
timeout=(3, 10))
dirty_response = None
for event in client:
if event.data == '"[DONE]"':
break;
dirty_response = event.data
if dirty_response is None:
self.log.error('bad session response from kagi')
return False
clean_response = self.clean_session_string(dirty_response)
self.speak(clean_response)
return True
except Exception as e:
self.log.error(f'error fetching kagi session response: {str(e)}')
return False
finally:
if client:
try:
client.close()
except:
pass
def get_using_api(self, message):
api_key = self.settings.get("kagiApiKey", "")
header_content = {
'Authorization': f'Bot {api_key}',
@ -36,48 +76,46 @@ class KagiSkill(FallbackSkill):
}
js_data = json.dumps({'query': message.data.get('utterance')})
read_url = 'https://kagi.com/api/v0/fastgpt'
response = requests.post(read_url,
data=js_data,
headers=header_content,
verify=False)
json_data = json.loads(response.text)
dirty_response = self.get_kagi_response(json_data)
if dirty_response is None:
self.speak("I wasn't able to retrieve an answer.")
else:
clean_response = self.clean_string(dirty_response)
try:
response = requests.post(read_url,
data=js_data,
headers=header_content,
verify=False)
json_data = json.loads(response.text)
dirty_response = self.get_api_response(json_data)
if dirty_response is None:
self.log.error('bad api response from kagi')
return False
clean_response = self.clean_api_string(dirty_response)
self.speak(clean_response)
return True
return True
except Exception as e:
self.log.error(f'error fetching kagi api response: {str(e)}')
return False
def clean_string(self, text):
# Remove asterisks and underscores
def clean_session_string(self, html_text):
pattern_refs = re.compile(r'<div[^>]*>.*$', re.DOTALL)
pattern_sups = re.compile(r'<sup[^>]*>.*?</sup>', re.DOTALL)
pattern_tags = re.compile(r'<.*?>')
text = re.sub(pattern_refs, '', html_text)
text = re.sub(pattern_sups, '', text)
text = re.sub(pattern_tags, '', text)
return html.unescape(text)
def clean_api_string(self, text):
text = text.replace('*', '').replace('_', '')
# Remove numbers in the brackets that Kagi uses for refs (e.g., 【1】【12】)
text = re.sub(r'\\d+】', '', text)
return text
def get_kagi_response(self, json_data):
try:
# Check if "data" exists and is a dict
if not isinstance(json_data.get("data"), dict):
self.log.error("'data' missing from kagi response")
return None
# Check if "output" exists in data
if "output" not in json_data["data"]:
self.log.error("'output' missing from kagi response")
return None
# Check if the output value is a string
output_value = json_data["data"]["output"]
if not isinstance(output_value, str):
self.log.error(f"'output' was not a string in kagi response (found {type(output_value).__name__})")
return None
return output_value
except Exception as e:
self.log.error(f"error parsing kagi response: {str(e)}")
def get_api_response(self, json_data):
if not isinstance(json_data.get("data"), dict):
self.log.error("'data' missing from kagi api response")
return None
if "output" not in json_data["data"]:
self.log.error("'output' missing from kagi api response")
return None
output_value = json_data["data"]["output"]
if not isinstance(output_value, str):
self.log.error(f"'output' was not a string in kagi api response (found {type(output_value).__name__})")
return None
return output_value

7
requirements.txt Normal file
View file

@ -0,0 +1,7 @@
certifi==2025.1.31
charset-normalizer==3.4.1
idna==3.10
requests==2.32.3
six==1.17.0
sseclient==0.0.27
urllib3==2.3.0