5iYS dZddlmZddlmZmZddlmZddlmZm Z ddl m Z m Z Gdde eZGd d e eZGd d e ZGd de ZGdde ZGdde ZGdde ZGdde ZGdde ZGdde ZGdde ZGdde ZGdd e ZGd!d"e ZGd#d$e ZGd%d&e ZGd'd(e ZGd)d*e ZGd+d,e Z Gd-d.e Z!y/)0zr Pydantic data models for the AI Video Production Pipeline. All data flows between agents use these typed models. ) annotations)datetimetimezone)Enum)AnyOptional) BaseModelFieldceZdZdZdZdZdZy)TargetPlatformtiktokinstagram_reelsyoutube_shortsyoutubeN)__name__ __module__ __qualname__TIKTOKINSTAGRAM_REELSYOUTUBE_SHORTSYOUTUBE//root/.openclaw/workspace/visionaryfx/models.pyr r s F'O%NGrr ceZdZdZdZdZy) PipelineModeplan plan_imagesgenerateN)rrrPLANPLAN_AND_IMAGESGENERATErrrrrs D#OHrrceZdZUdZedZded<edZded<eed Z d ed <eed Z d ed <eed Z d ed<edZ ded<edZ ded<edZded<edZded<edZded<edZded<eddZded <edd!Zded"<edd#Zded$<edd%Zded&<edd'Zded(<y))*CharacterTemplateuDImmutable character DNA – repeated verbatim in every image prompt.z+Short identifier for the character template descriptionstrnamez^Full physical description of the character. This is injected verbatim into every image prompt.physical_descriptionz-Visual style keywords applied to every promptdefault_factoryr& list[str]style_keywordsz0Keywords to explicitly exclude from every promptnegative_keywordszgAbsolute prohibitions - things that MUST NEVER appear in any image. More strict than negative_keywords.hard_prohibitionszMedium or medium-wide shot, eye-level or chest-level camera, no extreme angles, no dramatic lens changes, same framing logic across scenesdefaultcamera_settingsz_Real-world lighting matching the environment, natural shadows, photorealistic cinematic realismlighting_settingszThe character's POSE, BODY POSITION, and GESTURE MUST change per scene to match the script. DO NOT keep a neutral standing pose unless the script explicitly implies it. DO NOT repeat identical poses in consecutive scenes. pose_rulezInfer the environment directly from the script. Place the character naturally inside that environment. Environment must be realistic and context-appropriate. Vary environments across scenes for visual dynamism.environment_ruleaEAnalyze the script to determine contextually appropriate clothing. ONE clothing style is defined at the start and maintained consistently across ALL scenes unless the script explicitly implies a change. Clothing must be REALISTIC and contextually believable. Always describe clothing in FULL DETAIL (color, fabric type, fit). clothing_ruleucWhen the script mentions a BODY PART, ORGAN, HEALTH CONDITION, PAIN, or PHYSICAL SYMPTOM, that specific body area MUST be visually highlighted. Use distinct COLOR CODING: - PAIN/INFLAMMATION → Pulsing deep red or orange glow - FATIGUE/WEAKNESS → Faded blue-grey desaturation - BLOCKAGE/TENSION → Dense dark purple or black pressure cloud - DAMAGE/INJURY → Fractured visualization with amber-red highlight - HEALING/RECOVERY → Soft golden or green luminescent glow - NERVE SIGNAL/ACTIVITY → Electric blue or cyan pulse lines The highlight must look MEDICAL and SCIENTIFIC, not stylized or cartoonish.anatomical_highlight_ruleszhPhilosophy for showing consequences instead of symptoms. Maps script concepts to representative actions.r1r&consequence_philosophyz?Rules for death scenes: camera angle, posture, visual elements.death_scene_rulesz:Rules for depicting children: always abstract silhouettes. child_rulesz?Rules for infected/zombie characters: color + posture, no gore. zombie_rulesz>Rules for shot variety and intercalation throughout the video.shot_intercalation_rulesN)rrr__doc__r r(__annotations__r)listr-r.r/r2r3r4r5r6r7r:r;r<r=r>rrrr$r$$sN"OPD#P %=!#!&C!NI$)F$y$).$y !XOS#+saIs "Tc MM3 ',V ' #(:#C #UsPKUL#%*T%crr$ceZdZUeddZded<eddZded<ed d Zded <ed d dZded<edd dZ ded<ed d dZ ded<edd dZ ded<eddZ ded<eddZ ded<eddZded<ed d!Zded"<edd#d$d%&Zded'<ed(d)Zded*<ed+d,Zd-ed.<y+)/ VoiceSettings chatterboxz*TTS provider: 'chatterbox' or 'elevenlabs'r9r' tts_providerJBFqnCBsd6RMkjVDRZzbzElevenLabs voice IDvoice_ideleven_multilingual_v2zElevenLabs model IDmodel_idg?gg?)r1gelefloat stabilityg?similarity_booststyleg@speedptz5Language code for ChatterBox (e.g., 'pt', 'en', 'es')chatterbox_languager8z4Path to reference audio for voice cloning (optional)chatterbox_voice_refopenaizTTS model: openai or elevenlabspollinations_modelnovazPollinations voice namepollinations_voiceg?g@z Speech speedr1rJrKr&pollinations_speedmp3z,Audio format: mp3, opus, aac, flac, wav, pcmpollinations_formatNz"Optional API key for higher limits Optional[str]pollinations_api_key)rrrr rEr@rGrIrMrNrOrPrRrSrUrWrYr[r]rrrrCrCxsl@lmL#m"8F[\Hc\":H]^Hc^SSS9Iu9#DSSAeA5E555E55 %T?vww %b>t u#u$HBcdd#F@YZZ %cdsP^ __$U@noo*/Jn*o-orrCc~eZdZUedZded<edZded<edZded <ed Zded <y ) ExportSettings 1080x1920r0r' resolutionintfpsmp4 video_formatz9:16 aspect_ratioN) rrrr rar@rdrfrgrrrr_r_s>K0J0R C e,L#,f-L#-rr_ceZdZUdZedZded<edZded<ed Zded <ed Z ded <ed Z ded<eddZ ded<eddddZ ded<eddZ ded<y) SceneAnalysisz7Output of the Script Analyzer agent for a single scene.z%Sequential scene number starting at 1r%rc scene_numberz(Physical setting / location of the scener' environmentz8Character pose, body position and gesture for this scene pose_actionz"Emotional tone / mood of the scenemoodz0Exact narration text to be spoken for this scenenarration_textr8z/Optional delivery notes (pace, emotion, pauses)r9narration_notes zImportance score for coverage (1-10). Higher = more likely to get multi-angle coverage. 10=climax/action, 7-9=emotional intensity, 4-6=normal, 1-3=transition/expositionrXcoverage_priorityFzAI-determined: does this scene benefit from multi-angle coverage? True for action, emotion, climax. False for exposition, transition, static scenes.boolneeds_coverageN)rrrr?r rjr@rkrlrmrnrorsrurrrririsA*QRL#R)STKTNK"FGD#GFNC!EOS # [ s!]NDrriceZdZUdZedZded<edZded<ed Zd ed <ed d Z ded<y)SceneBreakdownz-Complete output of the Script Analyzer agent.zShort title for the videor%r'titlezTotal number of scenesrc total_sceneszOrdered list of sceneszlist[SceneAnalysis]scenesr8zBrief summary of the storyr9summaryN) rrrr?r rxr@ryrzr{rrrrwrwsH7#>?E3?*BCL#C"'4L"MF M1MNGSNrrwceZdZUdZedZded<edZded<edZd ed <ed Z d ed <ed Z d ed<edZ d ed<edZ d ed<edZ d ed<edZd ed<edZd ed<eddZd ed<eddZd ed<eddZd ed <ed!d"Zd#ed$<ed%d&Zded'<edd(Zd ed)<y!)*Framez5A single visual frame expanded from a narrative beat.z,Global sequential frame number starting at 1r%rc frame_numberz0Which narrative beat/scene this frame belongs to parent_scenezCamera shot type: wide shot, medium shot, medium close-up, close-up, extreme close-up, low angle, high angle, dutch angle, POV, over-the-shoulder, tracking shot, establishing shotr' camera_shotzfCamera movement: static, slow pan left/right, dolly forward/back, slight drift, tilt up/down, crane upcamera_movementznComposition: centered, rule of thirds left/right, foreground emphasis, background depth, symmetrical, diagonal compositionz)Specific environment focus for this frameenvironment_detailz1Specific character pose and action for this framerlz%Emotional tone of this specific framermz6What the viewer's eye should be drawn to in this frame visual_focusz6The narration text spoken during this frame's durationrnr8z)Delivery notes for this frame's narrationr9rozExact word or short phrase (1-5 words) from the script that this frame represents. This is the editing anchor - the moment in the audio where this image should cut in. sync_wordzIf the frame involves a body part, organ, or health condition, describe the anatomical highlight with color coding (e.g., 'lower back - red/orange glow for pain'). Leave empty if no anatomical highlight is needed.anatomical_highlightNzID grouping frames that show the same scene from different angles. Frames with the same coverage_group_id represent the same moment with different camera angles.r\coverage_group_idrzMIndex within coverage group: 0=master/wide, 1=medium, 2=close-up, 3+=variantscoverage_angle_indexzeType of coverage angle: 'master', 'medium', 'close-up', 'reverse', 'insert', or empty if not coveragecoverage_angle_type)rrrr?r r~r@rrrrrrlrmrrnrorrrrrrrrr}r}sy?*XYL#Y*\]L#]CK !COSGK$?GK"IJD#JLL# LNC!?OS_Is !&<!#(-i(} !&c!# %{ rr}c:eZdZUdZded<ded<ded<ded<y ) ExpandedFrameBreakdownu?Output of the FrameExpander agent – all frames for the video.r'rxr{rc total_framesz list[Frame]framesNrrrr?r@rrrrrsI J L rrcDeZdZUdZded<edZded<edZded <ed d Zded <ed d Z ded<ed d Z ded<ed d Z ded<ed d Z ded<ed d Z ded<ed d Zded<ed d Zded<y) ScenePromptsz7Output of the Prompt Engineer agent for a single frame.rcrjzwFull standalone image generation prompt with complete character description, environment, pose, camera, lighting, styler%r' image_promptzyImage-to-video motion prompt describing character action joint-by-joint with kinetic chain, timing, and muscle visibility video_promptr8zjExact word or short phrase (1-5 words) from the script - the editing anchor where this image should cut inr9rz_Body part + real anatomical structure (e.g., 'quadriceps expanding with visible muscle fibers')rz=Full clothing description for this scene (color, fabric, fit)clothing_descriptionz"Environment/setting for this scenerkz_ `#` %b>Y Z#Z8VWNCW(-W(}!&Q!# %R rrceZdZUdZded<ded<ded<ded<ded <ded <ded <d ed <ded<ded<eedZded<y)ProductionPlanuBThe complete production plan – the final output of the pipeline.r'rxr{rrLrrcrytarget_platformcharacter_template_namerCvoice_settingsr_export_settingszlist[ProductionScene]rzz5Step-by-step instructions to assemble the final videor*r,assembly_instructionsN)rrrr?r@r rArrrrrrsWL J L!!!!  !!## !!',K(9rrc>eZdZUdZedZded<eddZd ed <edd Zd ed <eddZ d ed<eddZ d ed<eddZ ded<eddZ ded<eddZ d ed<eddZd ed<eddZded<y) PipelineEventzAEvent emitted during pipeline execution for real-time UI updates.zeType of event: agent_start, agent_complete, agent_progress, image_generated, pipeline_complete, errorr%r' event_typeNzAgent number 1-5r9z Optional[int] agent_numberzHuman-readable agent namer\ agent_namezCurrent progress countprogress_currentzTotal items to processprogress_totalr8zHuman-readable status messagemessagezArbitrary event datazOptional[dict[str, Any]]datazBase64-encoded generated image image_base64zScene number for imageimage_scene_numbercdtjtjj S)N)rnowrutc isoformatrrrzPipelineEvent.s X\\ : D D FrzISO 8601 timestampr* timestamp)rrrr?r rr@rrrrrrrrrrrrrrsK4J#(BT"UL-U %d@[ \J \&+DF^&_m_$)$D\$]NM]1PQGSQ%*4E[%\D "\"'Bb"cL-c(-dH`(a aF(IsrrceZdZUdZej Zded<edZ ded<e jZ ded <d Z d ed <ee Zded<ee Zded<edZded<eddZded<eddZded<eddZded<edd Zd!ed"<ed#d$Zd%ed&<ed'd(d)d*+Zd!ed,<ed-d.Zded/<y )0PipelineConfigz'Runtime configuration for the pipeline.rmodez openai/gpt-4or0r' llm_modelr rNzOptional[CharacterTemplate]character_template)r+rCrr_routput output_dirg@z:Words per second for duration estimation (default 150 wpm)r9rLspeaking_rate_wpsg@z!Minimum scene duration in secondsmin_scene_durationg333333?z-Transition duration between scenes in secondstransition_durationz1Target number of image frames per minute of videorcframes_per_minuteFz?Generate multiple camera angles for each scene (coverage shots)rtenable_coverage_shotsrpz3Number of camera angles to generate per scene (2-5)rXcoverage_angles_per_sceneallzECoverage mode: 'all' (every scene) or 'smart' (AI decides key scenes) coverage_mode)rrrr?rr rr@r rr rrrrCrr_rrrrrrrrrrrrrrs&1%**D,*?3Is3&4&;&;O^;6:3:$)-$HNMH&+N&KO^KH-J-$P u!&7!"'C"#Gs #(U#4&+ I &s [M3rrN)"r? __future__rrrenumrtypingrrpydanticr r r'r rr$rCr_rirwr}rrrrrrrrrrrrrrrrs #' %S$3M MhpIp0.Y.I@OYO<I<~Y/9/d i y ) I  y &i&RY6I2)Y)r