Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fixed bug where named arguments did not match up in recurse function. #198

Merged
merged 3 commits into from
Sep 13, 2023
Merged

Fixed bug where named arguments did not match up in recurse function. #198

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
8ef9d7f
Fixed bug where named arguments did not match up in recurse function.
cmungall Sep 13, 2023
1036d3b
Move show_prompt parameter; name more args
caufieldjh Sep 13, 2023
cff3dac
Fix one more instance of incorrect param alignment
caufieldjh Sep 13, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion src/ontogpt/clients/openai_client.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,7 @@ def __post_init__(self):
self.api_key = get_apikey_value("openai")
openai.api_key = self.api_key

def complete(self, prompt, show_prompt: bool = False, max_tokens=3000, **kwargs) -> str:
def complete(self, prompt, max_tokens=3000, show_prompt: bool = False, **kwargs) -> str:
engine = self.model
logger.info(f"Complete: engine={engine}, prompt[{len(prompt)}]={prompt[0:100]}...")
if show_prompt:
Expand Down
8 changes: 4 additions & 4 deletions src/ontogpt/engines/gpt4all_engine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -80,7 +80,7 @@ def extract_from_text(
chunks = chunk_text(text, self.sentences_per_window)
extracted_object = None
for chunk in chunks:
raw_text = self._raw_extract(chunk, cls, object=object)
raw_text = self._raw_extract(chunk, cls=cls, object=object)
logging.info(f"RAW TEXT: {raw_text}")
if show_prompt:
logging.info(f" PROVIDED PROMPT:\n{self.last_prompt}")
Expand All @@ -97,11 +97,11 @@ def extract_from_text(
else:
extracted_object[k] = v
else:
raw_text = self._raw_extract(text, cls, object=object)
raw_text = self._raw_extract(text=text, cls=cls, object=object)
logging.info(f"RAW TEXT: {raw_text}")
if show_prompt:
logging.info(f" PROVIDED PROMPT:\n{self.last_prompt}")
extracted_object = self.parse_completion_payload(raw_text, cls, object=object)
extracted_object = self.parse_completion_payload(raw_text, cls=cls, object=object)
return ExtractionResult(
input_text=text,
raw_completion_output=raw_text,
Expand All @@ -111,7 +111,7 @@ def extract_from_text(
)

def _extract_from_text_to_dict(self, text: str, cls: ClassDefinition = None) -> RESPONSE_DICT:
raw_text = self._raw_extract(text, cls)
raw_text = self._raw_extract(text=text, cls=cls)
return self._parse_response_to_dict(raw_text, cls)

def iteratively_generate_and_extract(
Expand Down
23 changes: 14 additions & 9 deletions src/ontogpt/engines/spires_engine.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -59,9 +59,9 @@ class SPIRESEngine(KnowledgeEngine):
def extract_from_text(
self,
text: str,
show_prompt: bool = False,
cls: ClassDefinition = None,
object: OBJECT = None,
show_prompt: bool = False,
) -> ExtractionResult:
"""
Extract annotations from the given text.
Expand All @@ -75,7 +75,7 @@ def extract_from_text(
chunks = chunk_text(text, self.sentences_per_window)
extracted_object = None
for chunk in chunks:
raw_text = self._raw_extract(chunk, cls, show_prompt=show_prompt, object=object)
raw_text = self._raw_extract(chunk, cls=cls, object=object, show_prompt=show_prompt)
logging.info(f"RAW TEXT: {raw_text}")
next_object = self.parse_completion_payload(raw_text, cls, object=object)
if extracted_object is None:
Expand All @@ -90,7 +90,7 @@ def extract_from_text(
else:
extracted_object[k] = v
else:
raw_text = self._raw_extract(text=text, cls=cls, show_prompt=show_prompt, object=object)
raw_text = self._raw_extract(text=text, cls=cls, object=object, show_prompt=show_prompt)
logging.info(f"RAW TEXT: {raw_text}")
extracted_object = self.parse_completion_payload(raw_text, cls, object=object)
return ExtractionResult(
Expand All @@ -102,11 +102,11 @@ def extract_from_text(
)

def _extract_from_text_to_dict(self, text: str, cls: ClassDefinition = None) -> RESPONSE_DICT:
raw_text = self._raw_extract(text, cls, )
raw_text = self._raw_extract(text=text, cls=cls)
return self._parse_response_to_dict(raw_text, cls)

def generate_and_extract(
self, entity: str, show_prompt: bool = False, prompt_template: str = None, **kwargs
self, entity: str, prompt_template: str = None, show_prompt: bool = False, **kwargs
) -> ExtractionResult:
"""
Generate a description using GPT and then extract from it using SPIRES.
Expand Down Expand Up @@ -166,7 +166,7 @@ def _remove_parenthetical_context(s: str):
else:
curie = None
result = self.generate_and_extract(
next_entity, show_prompt=show_prompt, prompt_template=prompt_template, **kwargs
next_entity, prompt_template=prompt_template, show_prompt=show_prompt, **kwargs
)
if curie:
if result.extracted_object:
Expand Down Expand Up @@ -357,17 +357,17 @@ def _serialize_value(self, val: Any, slot: SlotDefinition) -> str:
return val

def _raw_extract(
self, text, show_prompt: bool = False, cls: ClassDefinition = None, object: OBJECT = None
self, text, cls: ClassDefinition = None, object: OBJECT = None, show_prompt: bool = False,
) -> str:
"""
Extract annotations from the given text.

:param text:
:return:
"""
prompt = self.get_completion_prompt(cls, text, object=object)
prompt = self.get_completion_prompt(cls=cls, text=text, object=object)
self.last_prompt = prompt
payload = self.client.complete(prompt, show_prompt)
payload = self.client.complete(prompt=prompt, show_prompt=show_prompt)
return payload

def get_completion_prompt(
Expand Down Expand Up @@ -465,11 +465,13 @@ def _parse_line_to_dict(
# The LLML may mutate the output format somewhat,
# randomly pluralizing or replacing spaces with underscores
field = field.lower().replace(" ", "_")
logging.debug(f" FIELD: {field}")
cls_slots = sv.class_slots(cls.name)
slot = None
if field in cls_slots:
slot = sv.induced_slot(field, cls.name)
else:
# TODO: check this
if field.endswith("s"):
field = field[:-1]
if field in cls_slots:
Expand Down Expand Up @@ -501,6 +503,7 @@ def _parse_line_to_dict(
transformed = False
slots_of_range = sv.class_slots(slot_range.name)
if self.recurse or len(slots_of_range) > 2:
logging.debug(f" RECURSING ON SLOT: {slot.name}, range={slot_range.name}")
vals = [self._extract_from_text_to_dict(v, slot_range) for v in vals]
else:
for sep in [" - ", ":", "/", "*", "-"]:
Expand Down Expand Up @@ -591,9 +594,11 @@ def ground_annotation_object(
if slot.range in self.schemaview.all_enums():
enum_def = self.schemaview.get_enum(slot.range)
new_ann[field] = []
logging.debug(f"FIELD: {field} SLOT: {slot.name}")
for val in vals:
if not val:
continue
logging.debug(f" VAL: {val}")
if isinstance(val, tuple):
# special case for pairs
sub_slots = sv.class_induced_slots(rng_cls.name)
Expand Down