How to implement a browser and editor for a directed graph database based on video in the form of an image sequence?

Question

Database

What is the most achievable way for a NOT programmer (implying some training, of course) to implement a browser editor for a directed graph database linking:
- video in the form of image sequences (movies - about 200,000 frames each; "stacked" in time - I guess so : time tree concept ) In the diagram - blue. In the database, they can be represented, of course, by links (paths) and not by actual files.
- a single multidimensional tree of criteria (criteria for evaluating a motion picture / video), which can be represented (for clarity of connection with the timeline) in a single dimension; in the ideal case, each node of which should potentially have a connection with each frame (frame) of the image sequence. The diagram is red. Concretization of the semantic component of the criteria: this tree of criteria is, in fact, the ontology of the cinematic sign (in its semiotic representation), one of the three levels of which is represented by the ontology of the cinematic form - the dominant level for this case. Therefore, we are talking primarily about the semantic annotation of the film image, as well as the analysis of the syntax (syntax) of the cinematic sign, and, ideally, pragmatics.
Those. such a database structure (simplified):
(The left half is made by direct links of the image with the criterion, and in the left half the links are represented by separate nodes (yellow); Each link (link-node))
high resolution

Features at least:
- add and edit nodes and links
- search in the database, scaling by time -lines of the sequence of images or its playback. Those. standardized database query.
- search in the database by tags (specific criteria, timecode) Additional
features:
- Changes in the structure of the tree of criteria
- Complex queries in the database, as it is implemented, for example, in neo4j: the shortest path, the number of neighboring but, etc. etc.
First of all, I am looking for a solution that is as simple as possible (for a NOT programmer) to implement, optionally with the possibility of further conversion into separate software or a plugin for NLE (Avid Media Composer, FCPX or Adobe Pemiere Pro, Davinci Resolve) in the future. The convenience of the design is not particularly worrying so far, so for simplicity I am looking for ready-made software on which I can implement this, because. there are no prospects for creating software from scratch yet.
From everything I've seen, the closest are: KeyLines and Gruff , allowing you to scroll time stamps. But I don’t understand at all how I can “attach” a simple image sequence player to these options. A good example of syncing video and text is "Video Transcription" on youtube.com. Partially referencing the question:What is simple graph database software?
There is little knowledge in UE4 blueprints and scanty Python Cinema4D scripting, initial acquaintance with neo4j.
Thank you for your advice!
PS: The practical significance of the tool: the analysis of film works (potentially any static or dynamic image in the context of film language), and when adding functionality - the development environment (design) of the "narrative media content" project. In addition to the described basic tasks, there are far-reaching ideas for adding such a tool: semantic search, synchronization with the script (such as ScriptSync), integration with the IMDb database and Kinopoisk, integration with 3D applications and more. While there are basic needs for functionality, I am working on the system development of a software project. Perhaps I will find like-minded people for implementation? Therefore, I mention the authors whose articles I saw on a similar topic on Habré: especially - Alexander Alekseev , mclander , @m1rko, OsipovRoman , AllexIn . I am ready and will be glad to have a substantive discussion on the topic. And I apologize if I disturbed respected authors in vain.