o oh*T@sddlZddlZddlZddlZddlZddlZddlmZddlm Z ddl m Z ddl m Z ddlmZmZddlmZddlmZdd lmZdd lmZmZdd lmZmZmZdd lmZdd lm Z mZddlm!Z!ddl"Z"ddl#m$Z$dZ%dZ&dZ'GdddeZ(GdddeZ)GdddeZ*GdddeZ+GdddeZ,GdddeZ-dd Z.d!d"Z/d#d$Z0d%d&Z1d7d(d)Z2d*d+Z3d,e4fd-d.Z5d/d0Z6d1d2Z7e8d3krd4Z9d5Z:d6Z;e7e9e:e;dSdS)8N)OllamaEmbeddings)FAISS)ChatGroq)ChatPromptTemplate)StrOutputParserPydanticOutputParser)OutputFixingParser)RunnablePassthrough) PdfReader) BaseModelField)ListOptionalDict)urlparse)Unionr)RunnableLambda) ChatOllama8gsk_CEh3itIpUAkEkEKsUDqVWGdyb3FYoTjqmXNTBHOSxJFK3obGTzXZznomic-embed-textz"44622834-f22df6f12cf45558ee180dd8dc@szeZdZUedZeed<edddZeeed<edddZ e eed <edd dZ eed <edd dZ eeed <dS) SlideContentflashtypeNz$An optional subheading for the slide description subheading.z(List of paragraphs for the slide content paragraphszUA specific and concise suggestion for a relevant visualization or image (max 5 words)visualization_suggestionzURL of the image for the slideimage) __name__ __module__ __qualname__r rstr__annotations__rrrr rrr#r#U/var/www/eduai.edurigo.com/storigo/production/generate_storigo_content_from_prompt.pyrs rc@s>eZdZUedddZeeefed<edddZ e ed<dS)StorigoContent1.7Dictionary of slide contents with slide numbers as keysrslides/Total token count for all the generated content token_countN rrr r r'rr!rr"r)intr#r#r#r$r%%s r%c@s^eZdZUedZeed<edddZeed<edddZe eed<edd dZ eed <d S) MCQContentQuestionr.zThe multiple-choice questionrquestionzA list of 4 answer optionsoptionsz0The correct answer (e.g., 'a', 'b', 'c', or 'd')correct_answerN) rrr r rr!r"r.r/r r0r#r#r#r$r,)s r,c@s6eZdZUedddZeeefed<dZ e ed<dS)StorigoContent.r&rr'rr)Nr*r#r#r#r$r1/s r1c@sFeZdZUedddZeeefed<edddZ eee fed<dS)StorigoContentMCQ.r&rr'z8Dictionary of MCQs with identifiers like 'mcq_1' as keysmcqsN) rrr r r'rr!rr"r3r,r#r#r#r$r25s  r2c@sFeZdZUedddZeeeee ffe d<edddZ e e d<dS)StorigoContentMCQMid.zYDictionary of slide contents with slide numbers as keys and MCQs with MCQ numbers as keysrr'r(r)N) rrr r r'rr!rrr,r"r)r+r#r#r#r$r4;s $r4c Cslz ttd}t|g|}d|}tj|dd|||WSty5}z tdt|d}~ww)N)modelzmy_embeddings/T)exist_okzError creating embeddings: ) r OLLAMA_MODELr from_textsosmakedirs save_local Exceptionr!)text client_id embeddingsvectors client_direr#r#r$create_embeddings@s   rCcst|}|dd|}td|}tgdfdd|D}|fdd|D}|dtdt|}d|S)N \w+ andortheaaninonattoforofwithbycg|]}|vr|qSr#r#.0word common_wordsr#r$ Xz)generate_search_query..crTr#r#rU)suggestion_wordsr#r$rZ\r[) extract_context_keywordsjoinrefindalllowersetsplitminlen)r slide_contentcontext_keywordscombined_querywordsfiltered_wordsprioritized_words query_wordsr#)rYr\r$generate_search_queryMs   rncs^|jpddd|j}td|}tgdfdd|D}tt|ddS)NrDrErFcrTr#r#rUrXr#r$rZir[z,extract_context_keywords..)rr_rr`rarbrclist)rgr=rjkeywordsr#rXr$r^cs  r^c Csd}t|dddddd}z/tj||d}||}|d r2t|d d d d d }|ddWStd|WdStjyX}ztdt|WYd}~dSd}~wt ys}ztdt|WYd}~dSd}~ww)Nzhttps://pixabay.com/api/photo horizontalr]true relevance)keyq image_type orientationper_page safesearchorder)paramshitscSs|d|dS)Nlikes downloadsr#xr#r#r$sz%fetch_pixabay_image..T)rwreverser webformatURLzNo image found for query: z#Error fetching image from Pixabay: z)Unexpected error in fetch_pixabay_image: ) PIXABAY_API_KEYrequestsgetraise_for_statusjsonsortedprintRequestExceptionr!r<)queryurlr~responsedata sorted_hitsrBr#r#r$fetch_pixabay_imagels6 rc Cs|stddSt|D]S}z1t||}td|dd|t|}|r2td||WStd|tdWq ty_}ztd|ddt|WYd}~q d}~wwtd |d dS) Nz%No visualization suggestion provided.zAttempt rz to fetch image for query: zValid image found: z!No image URL returned for query: z"Error in get_valid_image (attempt z): zNo valid image found after z attempts)rrangernrtimesleepr<r!)rrg max_attemptsattemptr image_urlrBr#r#r$get_valid_images&   (rcCstd|}t|S)NrE)r`rarf)r=tokensr#r#r$ count_tokenss rreturncCst|dr |j}nt|}|}d|vsd|vr3|d}|dd}|dkr3|dkr3|||}td d |}zt |}d |vrLt |d WSW|SY|S) NcontentzHere's another attemptz I apologize{}rrz '(\w+)\":z"\1": properties) hasattrrr!stripfindrfindr`subrloadsdumps) ai_messager=startendparsedr#r#r$quick_json_fixs&    rc2 szS|dkr tdtd}ntddd}d}ttdt|} d d d d fd d d | B|BB} | ||d} tt | j dd d} t dt | t d| j D]\} }t |jdkrq|jdt |jdksdqYt|tr||dk}nt|tr|dv}nd}|r| j D]*\} }|jrt|j|}|r||_qt d| dd|_qt d| dd|_qn | j D]}d|_qd}| j D]}|jdd|jd|j}|t|7}qt dt | t d| j }t ||rLt d d!}d"}ttd}t|}t|}i}t|}tdt |dD]z}t ||krg}t|t|dt |D]$}||}||}t |d#d$} t |d%g}!|| d&d|!q.cSr)N num_slidesr#rr#r#r$rrcsS)N)get_format_instructionsrparserr#r$rr)rrformat_instructions)rrcSr)Nrr#)itemr#r#r$rr)rwrrz2Additional information and context for this slide.r)1ruyes0Fz+Warning: No suitable image found for slide z after multiple attempts.z>https://via.placeholder.com/640x480.png?text=Placeholder+Imagez/Warning: No visualization suggestion for slide .rDresultSADASDaDASDDSADAz MCQ Starteda Based on the following context from the last two slides, generate one multiple-choice question (MCQ). The question should be relevant to the content and designed to test comprehension. **Context**: {context} The final output **must be valid JSON only**. Ensure all keys and string values are enclosed in double quotes and do not include any additional strings or explanations: Don't give any conversational like text, For example: "Here is the MCQ:". Just return JSON Output. Here is the pydantic class to validate the output that you give. Make sure the check always passes. class MCQContent(BaseModel): type: str = Field("Question") question: str = Field(..., description="The multiple-choice question") options: List[str] = Field(..., description="A list of 4 answer options") correct_answer: str = Field(..., description="The correct answer (e.g., 'a', 'b', 'c', or 'd')") CRITICAL: Return ONLY the JSON object with no additional text, explanations, or formatting. {format_instructions} JSON:u Based on the following context from the last two slides, generate one multiple-choice question (MCQ). The question should be relevant to the content and designed to test comprehension. Context: {context} The final output MUST be a **valid JSON object**, and nothing else. Do not include any explanation or conversational text (e.g., "Here is your MCQ" or "Let's try again"). Output format example (use this exact structure): {{ "type": "Question", "question": "What is the capital of France?", "options": ["Berlin", "London", "Paris", "Madrid"], "correct_answer": "Paris" }} Ensure: - All strings use double quotes - No trailing commas - Exactly 4 options - Output must be enclosed in ... tags - Only return the JSON block inside these tags - Never include any phrases like “I apologize”, “Here is”, or “Let’s retry”. Only return valid JSON. CRITICAL: Return ONLY the JSON object with no additional text, explanations, or formatting. rrorz:  context_text)llmr)contextrzV======================================================================================mcq_)r'r)z [Warning] Skipping missing MCQ: z Error generating slide content: )-r GROQ_API_KEYrrr1r from_templateinvokedictrr'itemsrrfrappend isinstancer+r!rrrrvaluesrr_rr,rqkeysrregetattrrrrrfrom_llmrr.r/r0 enumerater4r<)2rrnum_mcqsis_image is_questionquestion_positionGPUrslide_content_templateslide_content_promptslide_content_chainrordered_slides slide_keyrg is_image_boolrtoken_count_text text_content all_slides mcq_templatemcq_output_parser_template mcq_parser mcq_prompt!mcq_output_parser_template_promptr3 slide_keysicontext_slidesjrwslidetitleparasr json_fixeroutput_fixing_parser mcq_resultr)mcqinterleaved_content mcq_counter total_slides total_mcqsintervalidxmcq_keystorigo_contentstorigo_content_without_mcrBr#rr$"generate_slide_content_from_prompts                             rc Csft}d}d}d}d}tdt|||||||}t} tdt|td| |dd dS) NTrrStartingzGenerated Slide Content:z Process took z.2fz seconds)rrr) rrr start_timerrrrrgend_timer#r#r$mainsr__main__zCThe history and impact of artificial intelligence in modern societyF)r)s^