o h@sdZddlZddlmZmZddlmZmZmZdZ dddZ dd Z dd d Z d d Z GdddeZddZedkrAedSdS)z} A converter from constituency trees to dependency trees using CoreNLP's UniversalEnglish converter. ONLY works on English. N)DependencyConverterRequestDependencyConverterResponse) send_request build_treeJavaProtobufContextz8edu.stanford.nlp.trees.ProcessDependencyConverterRequestcCst|tt|dS)N classpath)rrCONVERTER_JAVA)requestrr ]/var/www/html/env_mimamsha/lib/python3.10/site-packages/stanza/server/dependency_converter.pysend_converter_request sr cCs*t}|jD] }|jt|jdq|S)zI Request format is simple: one tree per sentence in the document N)r sentencestreesappendr constituency)docr sentencer r r build_requests rcCs"t|}t||d}t||dS)zu Convert the constituency trees in the document, then attach the resulting dependencies to the sentences rN)rr attach_dependencies)rrr responser r r process_docs rc Cst|jt|jkrtdt|jt|jftt|j|jD]\}\}}|j}t|jt|jkrCtd|t|jt|jft|jt|j dkr`td|t|jdt|j ft t dt|jd}t }|j D]*}|j |vrtd|j |f| |j |j|j|j d_|j|j|j d_qr||} t| dksJ| D]} d|j| d_d|j| d_q|q!dS)Nz-Sent %d sentences but got back %d conversionszUSentence %d of the conversion should have %d words but got back %d nodes in the graphzsSentence %d of the conversion should have %d edges (one per word, plus the root) but got back %d edges in the graphz&Found two parents of %d in sentence %drroot)lenr conversions ValueError enumeratezipgraphwordsnodeedgesetrangetargetaddsourceheaddepdeprelbuild_dependencies) rrsent_idxr conversionrexpected_nodestargetsr"rootsrr r r r"s."    rcs*eZdZdZdfdd ZddZZS)DependencyConverterz Context window for the dependency converter This is a context window which keeps a process open. Should allow for multiple requests without launching new java processes each time. Ncstt||ttdSN)superr1__init__rr )selfr __class__r r r4HszDependencyConverter.__init__cCs t|}||}t|||S)zl Converts a constituency tree to dependency trees for each of the sentences in the document )rprocess_requestr)r5rr rr r r processKs  zDependencyConverter.processr2)__name__ __module__ __qualname____doc__r4r9 __classcell__r r r6r r1Asr1cCstjddd}|d}td|t|ddtd||d}td|tdd}||td|WddS1sGwYdS) Nenztokenize,pos,constituency) processorszI like blue antennae.z{:C}z $CLASSPATHrzAnd I cannot lie.)stanzaPipelineprintformatrr1r9)nlpr converterr r r mainTs   "rG__main__r2)r=rAstanza.protobufrr$stanza.server.java_protobuf_requestsrrrr r rrrr1rGr:r r r r s