Even though litellm type definitions and openai API specifies content as list of dictionaries (with type and relevant attributes potentially to allow multimodal inputs/outputs in addition to text), ollama has been demonstrating marshal errors. As a workaround this change simplifies the content as string when there is no more than one content part.
Fixes#642, #928, #376
PiperOrigin-RevId: 766890141
This is in response to the litellm v1.71.2 + ollama v0.9.0 sending function call chunks with 0 indices across multiple calls and lacking call ids.
Solutions introduced:
1. increment fallback index when accumulated arg becomes json parsable.
2. tolerate finish reason == stop when tool calls are present
3. fallback to index when tool call id is None
Fixes https://github.com/google/adk-python/issues/294
PiperOrigin-RevId: 766258344
--
8772b3de0b2fd04f246cc90c4c8032f9dc8cfed9 by Jaeyeon Kim <anencore94@gmail.com>:
fix: update unit test code for test_connection
- Fix for Unit Test Failures
--
8e0b45c2a64994bfda6401847e485fdd9edc8306 by Jaeyeon Kim <anencore94@gmail.com>:
fix useless changes
--
54efa004fa0fc9bcf78b49061221994650e162bc by Jaeyeon Kim <anencore94@gmail.com>:
fix conflict resolve issue
--
003ed4418c73b74bfba5ff055a364b62ddc18fa7 by Wei Sun (Jack) <weisun@google.com>:
Autoformat test_connections_client.py
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/578 from anencore94:bugfix/failed_unittests ba0e1d33ad905b0b046023b685c70f4ef05e4efa
PiperOrigin-RevId: 766221165
Copybara import of the project:
--
ffd6184d7e402b0787b0fa37fc09cd519adcc7f3 by Calvin Giles <calvin.giles@trademe.co.nz>:
fix: Call all tools in parallel calls during partial authentication
--
c71782a582ba825dbe2246cdb5be3f273ca90dca by seanzhou1023 <seanzhou1023@gmail.com>:
Update auth_preprocessor.py
--
843af6b1bc0bc6291cb9cb23acf11840098ba6dd by seanzhou1023 <seanzhou1023@gmail.com>:
Update test_functions_request_euc.py
--
955e3fa852420ecbf196583caa3cf86b7b80ab56 by seanzhou1023 <seanzhou1023@gmail.com>:
Update test_functions_request_euc.py
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/853 from calvingiles:fix-parallel-auth-tool-calls f44671e37b9fe44a25c9b1c0c25a26fc634b011c
PiperOrigin-RevId: 765639904
Right now the agent builder has to specify the bigquery scope explicitly for bigquery tools, which is somewhat unnecessary. With this change the user does not have to specify scopes, although they would still have the ability to overrides scopes if necessary.
PiperOrigin-RevId: 765354010
--
8baeb0b569eaedc638b20e46894178a3b878dbd6 by Amulya Bhatia <amulya.bhatia@t-online.de>:
test: unit tests for built_in_code_executor and unsafe_code_executor
--
cfac73b9271557ead96eb5fb419e05d88c6e8cd4 by Amulya Bhatia <amulya.bhatia@t-online.de>:
test: unit tests for built_in_code_executor and unsafe_code_executor
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/971 from iamulya:code-executor-tests 55290e27b5e58ef3835905aec88639e936318d01
PiperOrigin-RevId: 764976316
--
cef3ca1ed3493eebaeab3e03bdf5e56b35c0b8ef by Lucas Nobre <lucaas.sn@gmail.com>:
feat: Add index tracking to handle parallel tool call using litellm
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/759 from lucasnobre212:fix/issue_484 65e22934bf839f9ea03963b9ec6c23fdce03f59f
PiperOrigin-RevId: 764902433
Copybara import of the project:
--
79962881ca1c17eb6d7bd9dcf31a44df93c9badd by Almas Akchabayev <almas.akchabayev@gmail.com>:
fix: separate thinking from text parts in streaming mode
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/777 from almeynman:separate-thinking-and-text-parts-in-stream-mode b63dcc7fd0fc3973888dcbb9d4cc7e7e0a66e7f7
PiperOrigin-RevId: 764561932
When the user id contains special characters (i.e. an email), we have added in extra url parsing to address those characters. We have also added an if statement to use the correct url when there is no user_id supplied.
Copybara import of the project:
--
ef8499001afaea40bd037c4e9946b883e23a5854 by Danny Park <dpark@calicolabs.com>:
--
773cd2b50d15b9b056b47b6155df492b0ca8034c by Danny Park <dpark@calicolabs.com>:
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/996 from dpark27:fix/list_vertex_ai_sessions d351d7f6017c03165129adc7d0212f21d1340d88
PiperOrigin-RevId: 764522026
Copybara import of the project:
--
824b4379d59a375191f8ce10997efd4021d5d0b3 by Andres Videla <andres.videla@trademe.co.nz>:
Adding regex for claude-4 models to anthropic_llm and updating tests
--
8fa2a2df1931026dc803eee0e9b60e82e90c9efa by Wei Sun (Jack) <Jacksunwei@gmail.com>:
Adds trailing comma.
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/934 from avidelatm:feature/add-support-for-claude-4-models-to-anthropic_llm 8d10bacdbec952ec9832180ac6c1d220916641da
PiperOrigin-RevId: 764396694
Structures supported:
a) agents_dir/agent_name.py (with root_agent or agent.root_agent in it)
b) agents_dir/agent_name_folder/__init__.py (with root_agent or agent.root_agent in the package)
c) agents_dir/agent_name_folder/agent.py (where agent.py has root_agent)
PiperOrigin-RevId: 763943716
--
b781880d9bfb9786bd5e50314eaedc441fc2a93e by Stephen Smith <stephen.smith@newfront.com>:
Telemetry unit test for non-serializable data.
--
179da9db997bb3f992e126c9c64193ff7df67b3d by Stephen Smith <stephen.smith@newfront.com>:
When converting the llm_request to JSON, skip non-serializable data.
--
5dc68f4f5a6d12b753fdb81d1449716d13490afb by Stephen Smith <stephen.smith@newfront.com>:
Update _create_invocation_context() return type to InvocationContext.
--
23a33f754409fcd2a7641098d68cef7e4f1c72c6 by Stephen Smith <stephen.smith@newfront.com>:
Oops, remove unnecessary import.
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/879 from stephensmithnewfront:main f71e195a9ed157e4c0b3abfa74ae078b0c1a920a
PiperOrigin-RevId: 763922003
Copybara import of the project:
--
d01a8fd5f079bc4fca9e4b71796dbe65312ce9ff by Leo Yongsul Kim <ystory84@gmail.com>:
fix(DatabaseSessionService): Align event filtering and ordering logic
This commit addresses inconsistencies in how DatabaseSessionService
handles config.after_timestamp and config.num_recent_events
parameters, aligning its behavior with InMemorySessionService and
VertexAiSessionService.
Key changes:
- Made after_timestamp filtering inclusive
- Corrected num_recent_events behavior to fetch the N most recent events
- Refined timezone handling for after_timestamp
- Updated the unit test test_get_session_with_config to includeSessionServiceType.DATABASE, allowing verification of these fixes.
Fixes#911
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/915 from ystory:fix/database-session-timestamp-recency 5cc8cf5f5a5c0cb3e87f6ab178a5725d3f696c88
PiperOrigin-RevId: 763874840
--
73826d258b136f92a8da8171f7dc14d5f07de8dd by Calvin Giles <calvin.giles@trademe.co.nz>:
fix: Enable InMemoryRunner to be used in async tests
COPYBARA_INTEGRATE_REVIEW=https://github.com/google/adk-python/pull/868 from calvingiles:enable-test-runner-in-async fb9033ed6f350a3114859715cae51798f864ecf6
PiperOrigin-RevId: 763233472
