Mercurial Repositories > ollama Interface / file comparison
comparison: OllamaInterface/OllamaClient.py
OllamaInterface/OllamaClient.py
- changeset 5
- 6e8af43d537d
- parent 4
- 7dd1b9cd3150
- child 7
- eb1dec15b2f0
equal
deleted
inserted
replaced
| 4:7dd1b9cd3150 | 5:6e8af43d537d |
|---|---|
| 39 | 39 |
| 40 class OllamaClient(QObject): | 40 class OllamaClient(QObject): |
| 41 """ | 41 """ |
| 42 Class implementing the 'ollama' client. | 42 Class implementing the 'ollama' client. |
| 43 | 43 |
| 44 @signal replyReceived(content:str, role:str) emitted after a response from the | 44 @signal replyReceived(content:str, role:str, done:bool) emitted after a response |
| 45 'ollama' server was received | 45 from the 'ollama' server was received |
| 46 @signal modelsList(modelNames:list[str]) emitted after the list of model | 46 @signal modelsList(modelNames:list[str]) emitted after the list of model |
| 47 names was obtained from the 'ollama' server | 47 names was obtained from the 'ollama' server |
| 48 @signal detailedModelsList(models:list[dict]) emitted after the list of | 48 @signal detailedModelsList(models:list[dict]) emitted after the list of |
| 49 models was obtained from the 'ollama' server giving some model details | 49 models was obtained from the 'ollama' server giving some model details |
| 50 @signal runningModelsList(models:list[dict]) emitted after the list of | 50 @signal runningModelsList(models:list[dict]) emitted after the list of |
| 59 while processing the request | 59 while processing the request |
| 60 @signal serverStateChanged(ok:bool) emitted to indicate a change of the server | 60 @signal serverStateChanged(ok:bool) emitted to indicate a change of the server |
| 61 responsiveness | 61 responsiveness |
| 62 """ | 62 """ |
| 63 | 63 |
| 64 replyReceived = pyqtSignal(str, str) | 64 replyReceived = pyqtSignal(str, str, bool) |
| 65 modelsList = pyqtSignal(list) | 65 modelsList = pyqtSignal(list) |
| 66 detailedModelsList = pyqtSignal(list) | 66 detailedModelsList = pyqtSignal(list) |
| 67 runningModelsList = pyqtSignal(list) | 67 runningModelsList = pyqtSignal(list) |
| 68 pullStatus = pyqtSignal(str, str, int, int) | 68 pullStatus = pyqtSignal(str, str, int, int) |
| 69 serverVersion = pyqtSignal(str) | 69 serverVersion = pyqtSignal(str) |
| 99 self.__serverResponding = False # start with a faulty state | 99 self.__serverResponding = False # start with a faulty state |
| 100 | 100 |
| 101 self.__plugin.preferencesChanged.connect(self.__setHeartbeatTimer) | 101 self.__plugin.preferencesChanged.connect(self.__setHeartbeatTimer) |
| 102 self.__setHeartbeatTimer() | 102 self.__setHeartbeatTimer() |
| 103 | 103 |
| 104 def chat(self, model, messages): | 104 def chat(self, model, messages, streaming=True): |
| 105 """ | 105 """ |
| 106 Public method to request a chat completion from the 'ollama' server. | 106 Public method to request a chat completion from the 'ollama' server. |
| 107 | 107 |
| 108 @param model name of the model to be used | 108 @param model name of the model to be used |
| 109 @type str | 109 @type str |
| 110 @param messages list of message objects | 110 @param messages list of message objects |
| 111 @type list of dict | 111 @type list of dict |
| 112 """ | 112 @param streaming flag indicating to receive a streaming response |
| 113 # TODO: not implemented yet | 113 @type bool |
| 114 """ | |
| 114 ollamaRequest = { | 115 ollamaRequest = { |
| 115 "model": model, | 116 "model": model, |
| 116 "messages": messages, | 117 "messages": messages, |
| 118 "stream": streaming, | |
| 117 } | 119 } |
| 118 self.__sendRequest( | 120 self.__sendRequest( |
| 119 "chat", data=ollamaRequest, processResponse=self.__processChatResponse | 121 "chat", data=ollamaRequest, processResponse=self.__processChatResponse |
| 120 ) | 122 ) |
| 121 | 123 |
| 126 @param response dictionary containing the chat response | 128 @param response dictionary containing the chat response |
| 127 @type dict | 129 @type dict |
| 128 """ | 130 """ |
| 129 with contextlib.suppress(KeyError): | 131 with contextlib.suppress(KeyError): |
| 130 message = response["message"] | 132 message = response["message"] |
| 133 done = response["done"] | |
| 131 if message: | 134 if message: |
| 132 self.replyReceived.emit(message["content"], message["role"]) | 135 self.replyReceived.emit(message["content"], message["role"], done) |
| 133 | 136 |
| 134 def generate(self, model, prompt, suffix=None): | 137 def generate(self, model, prompt, suffix=None): |
| 135 """ | 138 """ |
| 136 Public method to request to generate a completion from the 'ollama' server. | 139 Public method to request to generate a completion from the 'ollama' server. |
| 137 | 140 |
| 140 @param prompt prompt to generate a response for | 143 @param prompt prompt to generate a response for |
| 141 @type str | 144 @type str |
| 142 @param suffix text after the model response (defaults to None) | 145 @param suffix text after the model response (defaults to None) |
| 143 @type str (optional) | 146 @type str (optional) |
| 144 """ | 147 """ |
| 145 # TODO: not implemented yet | |
| 146 ollamaRequest = { | 148 ollamaRequest = { |
| 147 "model": model, | 149 "model": model, |
| 148 "prompt": prompt, | 150 "prompt": prompt, |
| 149 } | 151 } |
| 150 if suffix is not None: | 152 if suffix is not None: |
| 161 | 163 |
| 162 @param response dictionary containing the generate response | 164 @param response dictionary containing the generate response |
| 163 @type dict | 165 @type dict |
| 164 """ | 166 """ |
| 165 with contextlib.suppress(KeyError): | 167 with contextlib.suppress(KeyError): |
| 166 self.replyReceived.emit(response["response"], "") | 168 self.replyReceived.emit(response["response"], "", response["done"]) |
| 167 | 169 |
| 168 def pull(self, model): | 170 def pull(self, model): |
| 169 """ | 171 """ |
| 170 Public method to ask the 'ollama' server to pull the given model. | 172 Public method to ask the 'ollama' server to pull the given model. |
| 171 | 173 |
