|
| 1 | +# coding=utf-8 |
| 2 | +""" |
| 3 | +@project: MaxKB |
| 4 | +@Author: 虎虎虎 |
| 5 | +@file: variable_splitting_node.py |
| 6 | +@desc: 变量拆分节点 |
| 7 | +""" |
| 8 | + |
| 9 | +import json |
| 10 | + |
| 11 | +from django.utils.translation import gettext_lazy as _ |
| 12 | +from jsonpath_ng.ext import parse |
| 13 | +from rest_framework import serializers |
| 14 | + |
| 15 | +from application.workflow.common import WorkflowType |
| 16 | +from application.workflow.i_node import INode |
| 17 | +from common.cache.mem_cache import MemCache |
| 18 | + |
| 19 | +jsonpath_expr_cache = MemCache( |
| 20 | + "parse_path", |
| 21 | + { |
| 22 | + "TIMEOUT": 3600, # 缓存有效期为 1 小时 |
| 23 | + "OPTIONS": { |
| 24 | + "MAX_ENTRIES": 1000, # 最多缓存 1000 个条目 |
| 25 | + "CULL_FREQUENCY": 10, # 达到上限时,删除约 1/10 的缓存 |
| 26 | + }, |
| 27 | + }, |
| 28 | +) |
| 29 | + |
| 30 | + |
| 31 | +class VariableSplittingNodeParamsSerializer(serializers.Serializer): |
| 32 | + input_variable = serializers.ListField(required=True, label=_("input variable")) |
| 33 | + variable_list = serializers.ListField(required=True, label=_("Split variables")) |
| 34 | + |
| 35 | + |
| 36 | +def parse_and_cache(path): |
| 37 | + jsonpath_expr = jsonpath_expr_cache.get(path) |
| 38 | + if not jsonpath_expr: |
| 39 | + jsonpath_expr = parse(path) |
| 40 | + jsonpath_expr_cache.set(path, jsonpath_expr) |
| 41 | + return jsonpath_expr |
| 42 | + |
| 43 | + |
| 44 | +def smart_jsonpath_search(data: dict, path: str): |
| 45 | + """智能 JSON Path 搜索。 |
| 46 | +
|
| 47 | + - 单个匹配: 直接返回值 |
| 48 | + - 多个匹配: 返回值的列表 |
| 49 | + - 无匹配: 返回 None |
| 50 | + """ |
| 51 | + jsonpath_expr = parse_and_cache(path) |
| 52 | + matches = jsonpath_expr.find(data) |
| 53 | + |
| 54 | + if not matches: |
| 55 | + return None |
| 56 | + elif len(matches) == 1: |
| 57 | + return matches[0].value |
| 58 | + else: |
| 59 | + return [match.value for match in matches] |
| 60 | + |
| 61 | + |
| 62 | +class VariableSplittingNode(INode): |
| 63 | + serializer_class = VariableSplittingNodeParamsSerializer |
| 64 | + supported_workflow_type_list = [ |
| 65 | + WorkflowType.APPLICATION, |
| 66 | + WorkflowType.KNOWLEDGE, |
| 67 | + WorkflowType.TOOL, |
| 68 | + ] |
| 69 | + type = "variable-splitting-node" |
| 70 | + |
| 71 | + def execute(self): |
| 72 | + node_params = self.get_parameters() |
| 73 | + |
| 74 | + input_variable = self.workflow_manage.get_reference_field( |
| 75 | + node_params.get("input_variable")[0], |
| 76 | + node_params.get("input_variable")[1:], |
| 77 | + ) |
| 78 | + variable_list = node_params.get("variable_list", []) |
| 79 | + |
| 80 | + if isinstance(input_variable, str): |
| 81 | + try: |
| 82 | + input_variable = json.loads(input_variable) |
| 83 | + except Exception: |
| 84 | + pass |
| 85 | + |
| 86 | + self.write_context("request", input_variable) |
| 87 | + response = {v["field"]: smart_jsonpath_search(input_variable, v["expression"]) for v in variable_list} |
| 88 | + self.write_context("result", response) |
| 89 | + for key, value in response.items(): |
| 90 | + self.write_context(key, value) |
| 91 | + |
| 92 | + def get_details(self, index: int = 0, position: dict = None, old_details: dict = None, **kwargs): |
| 93 | + details = super().get_details(index, position, old_details, **kwargs) |
| 94 | + details.update( |
| 95 | + { |
| 96 | + "request": self.get_context("request"), |
| 97 | + "result": self.get_context("result"), |
| 98 | + "status": self.status.value if self.status else None, |
| 99 | + } |
| 100 | + ) |
| 101 | + return details |
0 commit comments