U Kh n @sddlmZddlZddlmZddlmZddlmZddlZddl Z ddl m Z ddl Z ddl Z ddlTddlZddlmZdd lmZmZddlZd d d id d id did did did did did did did did dig iZdHddZdIddZddZddZd d!Zd"d#Zd$d%Zd&d'Zd(d)Zd*d+Zd,d-Z d.d/Z!d0d1Z"d2d3Z#d4d5Z$d6d7Z%d8d9Z&d:d;Z'dd?Z)d@dAZ*dBdCZ+dDdEZ,dFdGZ-dS)J)OpenAIN)tqdm)urlparse) HTTPBasicAuth)Image)*)fuzz)datetime timedelta$orZ categoryUriz dmoz/Businessz dmoz/Healthz dmoz/Homezdmoz/Recreationz dmoz/Societyzdmoz/Computersz dmoz/Shoppingz dmoz/Sportsz dmoz/Artsz dmoz/Games<cCst|||kS)a" Check if two titles are similar based on a similarity threshold. Args: title1 (str): First title. title2 (str): Second title. threshold (int): Similarity threshold (default 85%). Returns: bool: True if titles are similar, False otherwise. )rZtoken_sort_ratio)Ztitle1Ztitle2 thresholdr>/home/ankuromar296_gmail_com/publish_blogs_ai/gsc/post_news.pyis_similar_title6s rdc sddddiddiddiddiddidd idd idd igid |i||d giddgdddd}|dkr|dd}|dddigi||dd<t|t|}g} g} |j||dD]>} | ddrtfdd| Ds| | | q| S)a  Fetch news articles based on location and date range, filtering out similar titles. Args: location_uri (str): The URI of the location (e.g., country). date_start (str): The start date for fetching articles (format: YYYY-MM-DD). date_end (str): The end date for fetching articles (format: YYYY-MM-DD). max_items (int): Maximum number of articles to fetch. Default is 100. Returns: list: A list of unique or sufficiently distinct articles fetched based on the query. z$andr Z conceptUriz!http://en.wikipedia.org/wiki/Soilz:http://en.wikipedia.org/wiki/Geographic_information_systemz(http://en.wikipedia.org/wiki/Agriculturez$http://en.wikipedia.org/wiki/Farmingz%http://en.wikipedia.org/wiki/Forestryz'http://en.wikipedia.org/wiki/Fertilizerz)http://en.wikipedia.org/wiki/Agribusinessz)http://en.wikipedia.org/wiki/TraceabilityZ locationUri)Z dateStartZdateEndZblognewsZskipDuplicates)ZdataTypeZ isDuplicaterel)$queryz$filterZarticlesSortByenrlangeng)maxItemstitleNc3s|]}t|VqdSN)r).0Z seen_titlerrr sz$get_news_articles..)appendprintZQueryArticlesIterZinitWithComplexQueryZ execQuerygetany) erZ date_startZdate_end geographyr max_itemsqueryZ query_andqZ articles_listZ seen_titlesarticlerrrget_news_articlesEs|)1       r(cCsHtd|ddd}|}tdd|}tdd|d}|S) NNFKDasciiignorezutf-8z [^a-z0-9\s-]z\s+-) unicodedata normalizeencodedecodelowerresubstriprrrr create_slugsr6cCs,zt|}t|Wn|}YnX|Sr)ast literal_evalr)stringarrayrrrstring_to_arrays    r;cCsFd}d}t|d}|jjjddddd|dg||d }|jd jjS) N3sk-VHC3Gjk2iuFCPtANMrliT3BlbkFJ7wxsFMqRp4KreMhwLiWzsk-proj-O44Tus5LHWDwreXOqQOMjJqqKIVMrIYHNBoJSitbCH4OLdT5bDUp3Ey9n7qtt1zTsbwrUtHX6gT3BlbkFJLbzL1SHbiJDfiSin8Kyf--R9BfRQp4WTCa7kxGxQlZB-ALIqFlror4MCBBAcT5mc6k4a0T3PkAapi_keyzgpt-4o-mini-2024-07-18systemz:You are a expert in SEO and a representative of Farmonaut.rolecontentuser)modelmessages max_tokens temperaturer)rchat completionscreatechoicesmessagerC)promptrHrGr?client completionrrr call_openais  rQcCs6tjdd}|jjd||dd|dgd}|jdjS) Nzl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>zclaude-3-5-sonnet-20240620zYou are an SEO expert, a gis/ remote sensing expert, an agriculture and horticulture scientist, and a representative of Farmonaut (farmonaut.com).rDrA)rErGrHr@rFr) anthropic AnthropicrFrKrCtext)rNrHrGrOrMrrr call_genais rUcCs8d}d}t|d}|jjd|dddd}|jd j}|S) Nr<r=r>zdall-e-3 1024x1024standard)rErNsizequalitynr)rimagesgeneratedataurl)rNr?rOresponse image_urlrrr ai_image_urls  rbcCs2t|}|j}tj|d}|r.|ddSdS)NrXr,)rpathossplitext)r_ parsed_urlrcfile_extensionrrrget_file_extensionsrhc Cst|d}t|d}t|jd}|j|j}t||}|||ftj}|j|jdf}td|j d}| |||t ||} t|} | j dkr| | j } | |td|dS)NRGBAg?r)rrrrzWatermark added to )ropenconvertintheightwidthresizeLANCZOSnewrYpastealpha_compositemodesaver) main_image_pathwatermark_path main_image watermark new_height aspect_ratio new_widthposition transparentoutputoriginal_imagerrr add_watermark s       rc Cstj|r|ds"tdt|T}tj|d}|j dks\|j dkrfd|j krf| d}|d}| |d W5QRXt |td |d |d dS) Nz.pngz*The provided file is not a valid PNG file.r)riLAP transparencyRGBz.jpgJPEGz Converted z to z and removed the original PNG.)rdrcisfiler2endswith ValueErrorrrjrertinforkruremover) file_pathimg file_namejpg_pathrrrconvert_png_to_jpg0s     rc Csd}d}d}|d}t||}t|\}}|||d} |dtj|} t|d&} d| | |fi} tj|| || d } W5QRX| j d kr| d | d fSt d | j t d| j dSdS)Nhttps://www.farmonaut.com ankuromar296Tjat A2hz 9XMv pXJi YbV0 GR8oz/wp-json/wp/v2/media)alt_textcaption description_rbfile)filesauthjsonid source_urlz%Failed to upload media. Status code: z Response: )NN) r mimetypes guess_typerdrcbasenamerjrequestspost status_coderrrT)rrwp_url wp_username wp_passwordendpointr mime_typer media_data upload_namerrr`rrrupload_media_to_wordpressJs      rc Csd}d}d}|d}t||}z tj|d|d}|Wn8tjjk rx}ztdt|W5d}~XYnXtj|||d}|j d kr| Std |j dS) Nrrrz/wp-json/wp/v2/postsz/wp-json)rz Failed to access WordPress API: )rr)rzFailed to publish/update post: ) rrr raise_for_status exceptionsRequestException ExceptionstrrrrrT) post_datarrr posts_urlrr`errr publish_or_update_wordpress_postcs   $  rc Cs\g}d}d}|D]}|d}|dt|dtt|}ztj|dd}Wnttd}YnX|dkrvqtjtj |dd|j d krt |d &}|j d d D]} | r|| qW5QRXtd |t|dt|qtd|j qt|} | D]H}tj ||}t||\} } | r| r|d|| | dnqq|S)NrZinsta_files_newsrX/.T)stream)exist_okrwbi ) chunk_sizezFile downloaded successfully: z watermark.jpgz&Failed to download file. Status code: image)typerTrr_)rrhrr r traceback format_excrdmakedirsrcdirnamerrj iter_contentwriterrlistdirjoinrr) media_url_arrr media_info media_num folder_namer_rr`rchunk media_filesmedia_id media_urlrrrprocess_media2sD         rcCs|jdd}dgd}|dddd|jd|jdd t|d d D]V\}}d||d d |jd|jdd |d|dd }|d|qP|S)NrBreadcrumbList)@typeitemListElementrListItemrXHomez://)rr}nameitem)startr- ) rcr5splitrschemenetloc enumeratereplacerr)rf path_segments breadcrumbsr}segmentbreadcrumb_itemrrrcreate_breadcrumbs_fromURLs(  (rcCsPt|}dddddddgdf}tt|}td }d d |||||d S) N Organization Farmonautzhttps://farmonaut.comz"https://www.facebook.com/farmonautzhttps://twitter.com/farmonautz*https://www.linkedin.com/company/farmonautz#https://www.instagram.com/farmonaut)rrr_sameAsZzhttps://schema.orgZ NewsArticle)z@contextrheadlinerauthor breadcrumbZ datePublished)listrrr utcnow isoformat)rrblog_url image_urls organizationrZcurrent_timestamprrrcreate_news_schemas* rcCs4tt|d}d|d|d}t|ddd}|S)NzRewrite this text in roughly z tokens : Text: z< Note: Do not add any additional information whatsoever.rX@rHrG)rllenrQ)rrC total_tokensrNsummaryrrrget_news_summarys rcCs,d|d|d}t|ddd}t|}|S)NzFrom the content and summary provided, extract 5-10 SEO-optimized keywords that are highly relevant to the news topic: Title: z , Content: z| Output format: - Return only the refined array of keywords - Do not include any explanatory text or commentary rXrrUr;)rrCrNkeywordsrrr get_keywordssrcCs$d|d|d}t|ddd}|S)NzUsing the following keywords and summary, create a catchy SEO-optimized title for the news article. Make it 150 characters or less and include at least one power word. Keywords: z . Summary: z 4. Output format: - Return only the suggested title - Do not include any explanatory text, quotation marks, or additional commentaryrXrrrU)rrrNrrrrgenerate_titlesrcCs2d|d|d|d}t|ddd}t|}|S)NzUsing the following title, keywords, and summary, generate a set of 5-7 SEO-friendly keyphrases. Ensure the keyphrases are relevant and likely to rank well in search engines. Title: , Keywords: , Summary: ai 3. Output format: - Return results as an array of strings - Each keyphrase should be its own element in the array - Do not include any explanatory text or commentary - If fewer than 10 suitable keyphrases are found, include only those that meet the criteria Example output format: ["keyphrase 1", "keyphrase 2", "keyphrase 3", ...]rXrrr)rrrrN keyphrasesrrrget_keyphrases+s rcCs(d|d|d|d}t|dd}|S)Nz_Task: Generate an SEO-optimized news blog's meta description: Inputs: - Keywords: z - Key phrases: z - Reference Text: z Output format: - Provide only the 150-word summary - DO NOT OUTPUT ANY OTHER TEXT WITH THE RESPONSE rrr)rrrrNrrrgenerate_summary>s rc Csd|d|d|d|d }t|dd}tdt|}g}|D]:}z|t|td |WqBttYqBXqB|S) Nz Task: Suggest two SEO-optimizing AI image descriptions (DALL-E 3) for News Article. Inputs: a. Keywords: z b. Title: z c. Context: z d. KeyPhrases: a Output format: [ "Detailed description of first image, incorporating relevant keywords and focusing on a key aspect of the blog topic or Farmonaut's services.", "Detailed description of second image, highlighting a different facet of the blog content or Farmonaut's technology, using appropriate keywords." ] Note: Focus on creating descriptions that would result in images that add significant value to the blog post while optimizing for search engines. Ensure descriptions are distinct from each other and highly relevant to the content. DO NOT OUTPUT ANY OTHER TEXT WITH THE RESPONSE. rXrz Image Descriptions:z media url)rUrr;rrbrr)rrrrrNimage_descriptionsrimage_descriptionrrrgenerate_imagesPs( rcCsBd|d|d|d|d|d|d|d}t|d d d }|S) NzUsing the title, keywords, keyphrases, and summary, write a news article of upto 1500 words. Make sure the article is SEO-optimized, informative, engaging, and integrates the keyphrases naturally. Title: rz, Keyphrases: rz Key requirements: - Use HTML formatting including ,

,

, , ,
    ,
      ,
      , and

      tags where appropriate - Include and naturally incorporate as many of these keywords: z" - Include these key phrases: z - Write in the language of keywords/keywords if keywords/phrases are not in English - Mandatorily Localize the content if location names are available in keywords. - Add all the images from this JSON object: a with border-radius:16px, box-shadow: 10px 10px 15px, cursor: pointer. These images should open https://farmonaut.com/app_redirect when clicked. All images should be placed within 75% content of the blog. - Add links: - App Button Image: https://farmonaut.com/Images/web_app_button.png, Link on this button: https://farmonaut.com/app_redirect, height: 80px - API: https://sat.farmonaut.com/api - API Developer Docs:https://farmonaut.com/farmonaut-satellite-weather-api-developer-docs/ - Android App Button Image: https://farmonaut.com/wp-content/uploads/2020/01/get_it_on_google_play.png, Link on this Button: https://play.google.com/store/apps/details?id=com.farmonaut.android, height: 80px - iOS App Button Image: https://farmonaut.com/wp-content/uploads/2020/01/available_on_app_store.png, Link on this Button: https://apps.apple.com/in/app/farmonaut/id1489095847, height: 80px - Distribute these links within top 75% content of the blog with bold font - Use bullet points and subheadings (font color: #034d5c) to improve readability - Make the content mobile responsive Additional guidelines: - Achieve Flesch Reading Ease Score of at least 60 - Provide detailed explanations and examples - Ensure all content is factual and based on provided information - Organize information logically with clear transitions between sections - Use varied sentence structures and vocabulary for engaging reading - Mandatorily implement all the latest SEO guidelines provided by Google News Please generate the news article based on these requirements, ensuring it's well-structured, and upto 1500 words long. rXrrr)rrrrmediarN news_contentrrrgenerate_news_articleus" !r cCsL|dd}|dd}|}|}t||}t||}td|tdt|||}t|||}t||||} t| |} td|gd} | r| d| | r| d d nd| d <| d} t ||||| } td d t |}dd| D}t |||}dt |d}|| d<|| | d<d| d<|| d<d| d<dg| d<t| | S)NrzNo title availablebodyzNo content availablezGetting keyphraseszMedia Processed)rrrrrfeatured_mediaznews content processed zhttps://farmonaut.com/news/cSs g|]}|ddkr|dqS)rrr_r)rrrrr s z5generate_and_publish_news_article..z#rCpublishstatusexcerptrjcomment_statusi= categories)r rrrrrrrextendr r6rrdumpsr)r'Z title_articleZcontent_articleZ news_summaryrrrrrrrrrr rrZstructured_schemaZstructured_schema_scriptrrr!generate_and_publish_news_articlesB             rcCsjt|}|dkrtddSd|}t|ddD]4}zt|t|Wq0ttYq0Xq0dS)z Publish news articles spaced equally over 24 hours. Args: articles (list): List of articles to publish. generate_and_publish_news_article (func): Function to generate and publish an article. rzNo articles to publish.NiQzProcessing News articles)desc)rrrrtimesleeprr)articlesZtotal_articlesintervalr'rrrpublish_news_spacedsr)r )r).openairrRr urllib.parser requests.authrrr7PILrr3r. eventregistryr fuzzywuzzyrr r rrrr(r6r;rQrUrbrhrrrrrrrrrrrrrr rrrrrrs      %  ` '%3!% %'<