hueso/Alpaca
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Small CSS change and 'wait until Ollama loads to show Alpaca' fixed

Browse Source
This commit is contained in:
jeffser 2024-05-21 21:18:30 -06:00
parent fd86bc828d
commit d4a7576432
1 changed files with 8 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

11
src/window.py
View File

@ -789,7 +789,7 @@ class AlpacaWindow(Adw.ApplicationWindow):
margin_top = 6,
margin_start = 6,
margin_end = 6,
css_classes = ["card"]
css_classes = ["background"]
)
button_delete = Gtk.Button(
icon_name = "user-trash-symbolic",
@ -825,8 +825,13 @@ class AlpacaWindow(Adw.ApplicationWindow):
self.preferences_dialog.present(self)
def start_instance(self):
self.ollama_instance = subprocess.Popen(["/app/bin/ollama", "serve"], env={**os.environ, 'OLLAMA_HOST': f"127.0.0.1:{self.local_ollama_port}", "HOME": self.data_dir}, stdout=subprocess.PIPE)
sleep(5)
self.ollama_instance = subprocess.Popen(["/app/bin/ollama", "serve"], env={**os.environ, 'OLLAMA_HOST': f"127.0.0.1:{self.local_ollama_port}", "HOME": self.data_dir}, stdout=subprocess.PIPE, stderr=subprocess.PIPE, text=True)
while True:
err = self.ollama_instance.stderr.readline()
if err == '' and self.ollama_instance.poll() is not None:
break
if 'msg="inference compute"' in err: #Ollama outputs a line with this when it finishes loading, yeah
break
def stop_instance(self):
self.ollama_instance.kill()