BIDI_ML_INFO_EXTRACTION/BiddingKG/deepdive/app2.ddlog

@source
articles(
	@key
	@distributed_by
	id		text,
	@searchable
	content	text
).

#articles(
#	@key
#	id		text,
#	content	text
#).
#
#function articles over(
#	doc_id		text,
#	content	text
#) returns rows like articles
#	implementation "udf/articles.py" handles tsv lines.
#
#articles += articles(doc_id,content):-
#	articles(doc_id,content).
	
@source
sentences(
	@key
    @distributed_by
    doc_id         text,
    @key
    sentence_index int,
    @searchable
    sentence_text  text,
    tokens         text[],
    lemmas         text[],
    pos_tags       text[],
    ner_tags       text[],
    doc_offsets    int[],
    dep_types      text[],	
    dep_tokens     int[]
).

function nlp_markup over (
        doc_id  text,
        content text
    ) returns rows like sentences
    implementation "udf/nlp_markup_with_foolnltk.py" handles tsv lines.

sentences += nlp_markup(doc_id, content) :-
    articles(doc_id, content).


entity_mention(
	entity_id		text,
	entity_text		text,
	entity_type		text,
	doc_id			text,
	sentence_index	int,
	begin_index		int,
	end_index		int
).

function map_entity_mention over(
	doc_id			text,
	sentence_index	int,
	tokens			text[],
	ner_tags		text[]
)returns rows like entity_mention
implementation "udf/map_entity_mention.py" handles tsv lines.


entity_mention += map_entity_mention(
doc_id,sentence_index,tokens,ner_tags):-
sentences(doc_id,sentence_index,_,tokens,_,_,ner_tags,_,_,_).

relation(
	@key
	@distributed_by
	id		text,
	name	text,
	step_host	int,
	type_host	text,
	step_guest	int,
	type_guest	text
).

@source
feature_window(
	@key
	size	int
).

label_guest_BiddingAgency(
	entity_id		text,
	label			int,
	rule_id			text
).
function supervise_guest_BiddingAgency over(
	entity_id	text,entity_begin int,entity_end int,
	doc_id	text,
	sentence_index	int,
	sentence_text	text,
	tokens	text[],
	pos_tags	text[],
	ner_tags	text[]
)returns rows like label_guest_BiddingAgency
implementation "/udf/supervise_guest_BiddingAgency.py" handles tsv lines.

label_guest_BiddingAgency += supervise_guest_BiddingAgency(
entity_id,entity_begin,entity_end,doc_id,sentence_index,sentence_text,tokens,pos_tags,ner_tags
):-
entity_mention(entity_id,_,guest_type,doc_id,sentence_index,entity_begin,entity_end),
relation(id,_,_,_,_,guest_type),
id = "BiddingAgency",
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_).


label_guest_Tenderer(
	entity_id		text,
	label			int,
	rule_id			text
).
function supervice_guest_Tenderer over(
	entity_id	text,entity_begin int,entity_end int,
	doc_id	text,
	sentence_index	int,
	sentence_text	text,
	tokens			text[],
	pos_tags		text[],
	ner_tags		text[]
)returns rows like label_guest_Tenderer
implementation "udf/supervise_guest_Tenderer.py" handles tsv lines.

label_guest_Tenderer+= supervice_guest_Tenderer(
entity_id,entity_begin,entity_end,
doc_id,
sentence_index,
sentence_text,
tokens,
pos_tags,
ner_tags
):-
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_),
entity_mention(entity_id,_,guest_type,doc_id,sentence_index,entity_begin,entity_end),
relation(id,_,_,_,_,guest_type),
id="Tenderer".


label_guest_SignUpStart(
	entity_id		text,
	label			int,
	rule_id			text
).
function supervice_guest_SignUpStart over(
	entity_id	text,entity_begin int,entity_end int,
	doc_id	text,
	sentence_index	int,
	sentence_text	text,
	tokens			text[],
	pos_tags		text[],
	ner_tags		text[]
)returns rows like label_guest_SignUpStart
implementation "udf/supervise_guest_SignUpStart.py" handles tsv lines.

label_guest_SignUpStart+= supervice_guest_SignUpStart(
entity_id,entity_begin,entity_end,
doc_id,
sentence_index,
sentence_text,
tokens,
pos_tags,
ner_tags
):-
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_),
entity_mention(entity_id,_,guest_type,doc_id,sentence_index,entity_begin,entity_end),
relation(id,_,_,_,_,guest_type),
id="SignUpStart".

label_guest_BiddingOT(
	entity_id		text,
	label			int,
	rule_id			text
).
function supervice_guest_BiddingOT over(
	entity_id	text,entity_begin int,entity_end int,
	doc_id	text,
	sentence_index	int,
	sentence_text	text,
	tokens			text[],
	pos_tags		text[],
	ner_tags		text[]
)returns rows like label_guest_BiddingOT
implementation "udf/supervise_guest_BiddingOT.py" handles tsv lines.

label_guest_BiddingOT+= supervice_guest_BiddingOT(
entity_id,entity_begin,entity_end,
doc_id,
sentence_index,
sentence_text,
tokens,
pos_tags,
ner_tags
):-
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_),
entity_mention(entity_id,_,guest_type,doc_id,sentence_index,entity_begin,entity_end),
relation(id,_,_,_,_,guest_type),
id="BiddingOT".


label_guest_SignUpEnd(
	entity_id		text,
	label			int,
	rule_id			text
).
function supervice_guest_SignUpEnd over(
	entity_id	text,entity_begin int,entity_end int,
	doc_id	text,
	sentence_index	int,
	sentence_text	text,
	tokens			text[],
	pos_tags		text[],
	ner_tags		text[]
)returns rows like label_guest_SignUpEnd
implementation "udf/supervise_guest_SignUpEnd.py" handles tsv lines.

label_guest_SignUpEnd+= supervice_guest_SignUpEnd(
entity_id,entity_begin,entity_end,
doc_id,
sentence_index,
sentence_text,
tokens,
pos_tags,
ner_tags
):-
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_),
entity_mention(entity_id,_,guest_type,doc_id,sentence_index,entity_begin,entity_end),
relation(id,_,_,_,_,guest_type),
id="SignUpEnd".

label_guest_WinnerOfBidding(
	entity_id		text,
	label			int,
	rule_id			text
).
function supervice_guest_WinnerOfBidding over(
	entity_id	text,entity_begin int,entity_end int,
	doc_id	text,
	sentence_index	int,
	sentence_text	text,
	tokens			text[],
	pos_tags		text[],
	ner_tags		text[]
)returns rows like label_guest_WinnerOfBidding
implementation "udf/supervise_guest_WinnerOfBidding.py" handles tsv lines.

label_guest_WinnerOfBidding += supervice_guest_WinnerOfBidding(
entity_id,entity_begin,entity_end,
doc_id,
sentence_index,
sentence_text,
tokens,
pos_tags,
ner_tags
):-
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_),
entity_mention(entity_id,_,guest_type,doc_id,sentence_index,entity_begin,entity_end),
relation(id,_,_,_,_,guest_type),
id="WinnerOfBidding".

label_guest_FirstCandidate(
	entity_id		text,
	label			int,
	rule_id			text
).
function supervice_guest_FirstCandidate over(
	entity_id	text,entity_begin int,entity_end int,
	doc_id	text,
	sentence_index	int,
	sentence_text	text,
	tokens			text[],
	pos_tags		text[],
	ner_tags		text[]
)returns rows like label_guest_FirstCandidate
implementation "udf/supervise_guest_FirstCandidate.py" handles tsv lines.

label_guest_FirstCandidate += supervice_guest_FirstCandidate(
entity_id,entity_begin,entity_end,
doc_id,
sentence_index,
sentence_text,
tokens,
pos_tags,
ner_tags
):-
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_),
entity_mention(entity_id,_,guest_type,doc_id,sentence_index,entity_begin,entity_end),
relation(id,_,_,_,_,guest_type),
id="FirstCandidate".

label_guest_SecondCandidate(
	entity_id		text,
	label			int,
	rule_id			text
).
function supervice_guest_SecondCandidate over(
	entity_id	text,entity_begin int,entity_end int,
	doc_id	text,
	sentence_index	int,
	sentence_text	text,
	tokens			text[],
	pos_tags		text[],
	ner_tags		text[]
)returns rows like label_guest_SecondCandidate
implementation "udf/supervise_guest_SecondCandidate.py" handles tsv lines.

label_guest_SecondCandidate += supervice_guest_SecondCandidate(
entity_id,entity_begin,entity_end,
doc_id,
sentence_index,
sentence_text,
tokens,
pos_tags,
ner_tags
):-
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_),
entity_mention(entity_id,_,guest_type,doc_id,sentence_index,entity_begin,entity_end),
relation(id,_,_,_,_,guest_type),
id="SecondCandidate".

label_guest_ThirdCandidate(
	entity_id		text,
	label			int,
	rule_id			text
).
function supervice_guest_ThirdCandidate over(
	entity_id	text,entity_begin int,entity_end int,
	doc_id	text,
	sentence_index	int,
	sentence_text	text,
	tokens			text[],
	pos_tags		text[],
	ner_tags		text[]
)returns rows like label_guest_ThirdCandidate
implementation "udf/supervise_guest_ThirdCandidate.py" handles tsv lines.

label_guest_ThirdCandidate += supervice_guest_ThirdCandidate(
entity_id,entity_begin,entity_end,
doc_id,
sentence_index,
sentence_text,
tokens,
pos_tags,
ner_tags
):-
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_),
entity_mention(entity_id,_,guest_type,doc_id,sentence_index,entity_begin,entity_end),
relation(id,_,_,_,_,guest_type),
id="ThirdCandidate".



feature_entity(
	@key
	@references(relation="isBiddingAgency",column="entity_id",alias="isBiddingAgency")
	entity_id	text,
	feature		text,
	window_size	int
).
function feature_entity over(
	entity_id	text,
	entity_begin	int,
	entity_end		int,
	doc_id		text,
	sentence_index	int,
	tokens		text[],
	pos_tags	text[],
	ner_tags	text[],
	window_size		int
)returns rows like feature_guest_BiddingAgency
implementation "/udf/feature_entity.py" handles tsv lines.

feature_entity += feature_entity(
	entity_id,entity_begin,entity_end,doc_id,sentence_index,tokens,pos_tags,ner_tags,window_size
):-
entity_mention(entity_id,entity_text,_,doc_id,sentence_index,entity_begin,entity_end),
sentences(doc_id,sentence_index,sentence_text,tokens,_,pos_tags,ner_tags,_,_,_),
feature_window(window_size).



@extraction
is_biddingagency?(
	@key
	@references(relation="entity_mention",column="entity_id",alias="entity_mention")
	entity_id	text
).

is_biddingagency(entity_id) = if label>0 then TRUE else if label<0 then FALSE else NULL end:-
label_guest_BiddingAgency(entity_id,label,_).

@weight(f_BiddingAgency)
is_biddingagency(entity_id):-
feature_entity(entity_id,f_BiddingAgency,window),
label_guest_BiddingAgency(entity_id,_,_),
window=10.

@extraction
is_winnerofbidding?(
	@key
	@references(relation="entity_mention",column="entity_id",alias="entity_mention")
	entity_id	text
).

is_winnerofbidding(entity_id) = if label>0 then TRUE else if label<0 then FALSE else NULL end:-
label_guest_WinnerOfBidding(entity_id,label,_).

@weight(f_WinnerOfBidding)
is_winnerofbidding(entity_id):-
feature_entity(entity_id,f_WinnerOfBidding,window),
label_guest_WinnerOfBidding(entity_id,_,_),
window=10.


candidate_entity_pair(
	p1_id	text,
	p1_name	text,
	p2_id	text,
	p2_name	text
).

num_entity(doc_id,sentence_index,COUNT(p)):-
entity_mention(p,_,_,doc_id,sentence_index,_,_).

function map_candidate_entity_pair over(
	entity1_id		text,
	entity1_name		text,
	entity2_id		text,
	entity2_name		text
)returns rows like candidate_entity_pair
implementation "udf/map_candidate_entity_pair.py" handles tsv lines.

candidate_entity_pair += map_candidate_entity_pair(entity1,entity1_name,entity2,entity2_name):-
num_entity(same_doc,same_sentence,num_p),
entity_mention(entity1,entity1_name,_,same_doc,same_sentence,entity1_begin,_),
entity_mention(entity2,entity2_name,_,same_doc,same_sentence,entity2_begin,_),
num_p<5,
entity1!=entity2,
entity1_name!=entity2_name,
entity1_begin!=entity2_begin.


feature_entity_pair(
	@key
	@references(relation="has_entrust",column="entity1_id",alias="has_entrust")
	entity1_id	text,
	@key
	@references(relation="has_entrust",column="entity2_id",alias="has_entrust")
	entity2_id	text,
	feature		text,
	window_size	int
).

function feature_entity_pair over(
	entity1_id		text,
	entity2_id		text,
	entity1_begin_index	text,
	entity1_end_index	text,
	entity2_begin_index	text,
	entity2_end_index	text,
	doc_id			text,
	sentence_index	int,
	tokens			text[],
	lemmas			text[],
	pos_tags		text[],
	ner_tags		text[],
	dep_types		text[],
	dep_tokens		text[],
	window_size		int
)returns rows like feature_entity_pair
implementation "udf/feature_entity_pair.py" handles tsv lines.

feature_entity_pair += feature_entity_pair(
entity1_id,entity2_id,entity1_begin_index,entity1_end_index,entity2_begin_index,entity2_end_index,
doc_id,sentence_index,tokens,lemmas,pos_tags,ner_tags,dep_types,dep_tokens,window_size):-
sentences(doc_id,sentence_index,_,tokens,lemmas,pos_tags,ner_tags,_,dep_types,dep_tokens),
entity_mention(entity1_id,_,_,doc_id,sentence_index,entity1_begin_index,entity1_end_index),
entity_mention(entity2_id,_,_,doc_id,sentence_index,entity2_begin_index,entity2_end_index),
entity1_id!=entity2_id,
feature_window(window_size),
window_size=10.


@extraction
label_entrust(
	@key
	@references(relation="has_entrust",column="entity1_id",alias="has_entrust")
	entity1_id	text,
	@key
	@references(relation="has_entrust",column="entity2_id",alias="has_entrust")
	entity2_id	text,
	@navigable
	label		int,
	@navigable
	rule_id		text
).

label_entrust(entity1_id,entity2_id,0,NULL):-candidate_entity_pair(entity1_id,_,entity2_id,_).

function supervise_entrust over(
	entity1_id 	text,	entity1_begin 	int,	entity1_end int,
	entity2_id	text,	entity2_begin	int,	entity2_end int,
	doc_id		text,
	sentence_index int,
	sentence_text	text,
	tokens			text[],
	lemmas			text[],
	pos_tags		text[],
	ner_tags		text[],
	dep_types		text[],
	dep_tokens		int[]
)returns(
	entity1_id text,entity2_id text,label int,rult_id text 
)
implementation "udf/supervise_entrust.py" handles tsv lines.

label_entrust += supervise_entrust(
entity1_id,entity1_begin,entity1_end,
entity2_id,entity2_begin,entity2_end,
doc_id,sentence_index,sentence_text,
tokens,lemmas,pos_tags,ner_tags,dep_types,dep_tokens
):-
candidate_entity_pair(entity1_id,_,entity2_id,_),
entity_mention(entity1_id,entity1_name,_,doc_id,sentence_index,entity1_begin,entity1_end),
entity_mention(entity2_id,entity2_name,_,_,_,entity2_begin,entity2_end),
sentences(
	doc_id,sentence_index,sentence_text,
	tokens,lemmas,pos_tags,ner_tags,_,dep_types,dep_tokens
).

label_entrust_resolved(entity1_id,entity2_id,SUM(vote)):-label_entrust(entity1_id,entity2_id,vote,rule_id).


@extraction
has_entrust?(
	@key
    @references(relation="entity_mention", column="entity_id", alias="entity_mention")
	entity1_id	text,
	@key
    @references(relation="entity_mention", column="entity_id", alias="entity_mention")
	entity2_id	text
).

has_entrust(entity1_id,entity2_id) = if label>0 then TRUE else if label<0 then FALSE else NULL end :- 
label_entrust_resolved(entity1_id,entity2_id,label).

@weight(f)
has_entrust(entity1_id,entity2_id):-
	feature_entity_pair(entity1_id,entity2_id,f,window),
	window=10.