feat: add agent file processing

[maksymbuleshnyi]

Important

Enhance FileReadTool with intelligent file type detection and content extraction using converters and LLMs for various file types.

• Behavior:
  • FileReadTool in file_tools.py now detects file types and uses converters for text extraction.
  • Supports PDF, DOCX, PPTX, HTML, TXT, IMAGE, and more.
  • For large files, returns chunked content.
  • Uses LLM for image and PDF processing if supported.
• Converters:
  • Adds detect_file_type() function in file_tools.py for file type detection.
  • Introduces DEFAULT_FILE_TYPE_TO_CONVERTER_CLASS_MAP for mapping file types to converters.
  • LLMImageConverter and LLMPDFConverter in llm_text_extractor.py handle image and PDF text extraction.
• File Types:
  • New FileType enum and EXTENSION_MAP in file_types.py for file type management.
• Tests:
  • Updates test_file_tools.py to test new file reading and writing functionalities.

^{This description was created by for eaeafec. You can customize this summary. It will automatically update as commits are pushed.}

[ellipsis-dev]

Caution

Changes requested ❌

Reviewed everything up to 3201378 in 2 minutes and 9 seconds. Click for details.

• Reviewed 566 lines of code in 3 files
• Skipped 0 files when reviewing.
• Skipped posting 5 draft comments. View those below.
• Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. dynamiq/nodes/converters/llm_text_extractor.py:376

• Draft comment:
  Ensure run_kwargs is a dict and Python>=3.9 is assumed, as the dict union operator (|) is used.
• Reason this comment was not posted:
  Comment was not on a location in the diff, so it can't be submitted as a review comment.

2. dynamiq/utils/file_types.py:22

• Draft comment:
  Ensure EXTENSION_MAP is kept updated as new file types and converters are added.
• Reason this comment was not posted:
  Comment did not seem useful. Confidence is useful = 0% <= threshold 50% This comment is asking the PR author to ensure that EXTENSION_MAP is kept updated, which is a general reminder rather than a specific code suggestion or issue. It doesn't point out a specific problem or suggest a specific improvement.

3. dynamiq/nodes/tools/file_tools.py:223

• Draft comment:
  Typographical error: The print statement outputs 'starting converton'. Did you mean 'starting conversion' or 'starting converter'? Please correct the spelling.
• Reason this comment was not posted:
  Decided after close inspection that this draft comment was likely wrong and/or not actionable: usefulness confidence = 10% vs. threshold = 50% Print statements are typically used for debugging and should generally be removed before code is merged. The typo is in a debug print statement that likely shouldn't be in the final code anyway. The comment is technically correct but focuses on the wrong issue - instead of fixing the typo, the print statements should probably be removed entirely or replaced with proper logging. Maybe the print statements are intentionally left in for ongoing debugging purposes? Maybe the typo could cause confusion for other developers? Even if the prints are intentionally left in, fixing a typo in debug prints is extremely low value. The code already has proper logging statements elsewhere. The comment should be removed as it focuses on an unimportant typo in what appears to be temporary debug code.

4. dynamiq/nodes/tools/file_tools.py:233

• Draft comment:
  Typographical/placeholder text: The print message 'Asdasdasd' appears to be a placeholder or typo. Consider removing or correcting it.
• Reason this comment was not posted:
  Marked as duplicate.

5. dynamiq/nodes/tools/file_tools.py:364

• Draft comment:
  Typo/Concatenation: In the log warning message for no extracted text content, the two string literals are concatenated as "no text content extracted from file," + "falling back to raw content" resulting in "file,falling" without a space. Consider adding a leading space (e.g., " falling back to raw content") to ensure proper spacing in the log message.
• Reason this comment was not posted:
  Decided after close inspection that this draft comment was likely wrong and/or not actionable: usefulness confidence = 10% vs. threshold = 50% While technically correct that there's a missing space, this is an extremely minor issue that only affects a log message. Log messages are for debugging purposes and a missing space doesn't impact functionality. This feels like the kind of nitpicky comment that adds noise without real value. The missing space could make the log message slightly harder to read for developers debugging issues. The fix is trivial and would improve readability. While true that it would be slightly more readable, this is an extremely low-priority cosmetic issue in debug output. The message is still perfectly understandable. This comment should be deleted as it's too minor of an issue to warrant a PR comment. Log message formatting is not important enough to flag in code review.

Workflow ID: wflow_Xh3Fs8yff5m2G6Qj

^{You can customize by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

[ellipsis-dev]

Important

Looks good to me! 👍

Reviewed f17cdbe in 1 minute and 3 seconds. Click for details.

• Reviewed 38 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 3 draft comments. View those below.
• Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. dynamiq/nodes/tools/file_tools.py:222

• Draft comment:
  Removed debug print statements ('print("starting converton ")' and 'print(detected_type)'). Using logging is preferred over print for debugging.
• Reason this comment was not posted:
  Comment looked like it was already resolved.

2. dynamiq/nodes/tools/file_tools.py:240

• Draft comment:
  Removed debug print of merged kwargs. Consider using logging.debug if internal state info is needed.
• Reason this comment was not posted:
  Comment was not on a location in the diff, so it can't be submitted as a review comment.

3. dynamiq/nodes/tools/file_tools.py:338

• Draft comment:
  Removed debug printing of text_content in execute. Logging should be used instead of print statements for production.
• Reason this comment was not posted:
  Comment looked like it was already resolved.

Workflow ID: wflow_vzsJwfO8BApS3V5B

^{You can customize by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

[ellipsis-dev]

Important

Looks good to me! 👍

Reviewed 1ba617e in 1 minute and 4 seconds. Click for details.

• Reviewed 82 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 5 draft comments. View those below.
• Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. dynamiq/nodes/tools/file_tools.py:2

• Draft comment:
  Good: Added 'import os' for path operations.
• Reason this comment was not posted:
  Confidence changes required: 0% <= threshold 50% None

2. dynamiq/nodes/tools/file_tools.py:47

• Draft comment:
  Using os.path.splitext for file extension extraction is more robust than manual splitting.
• Reason this comment was not posted:
  Confidence changes required: 0% <= threshold 50% None

3. dynamiq/nodes/tools/file_tools.py:163

• Draft comment:
  Removed redundant inline comments in _setup_converters improves code clarity.
• Reason this comment was not posted:
  Confidence changes required: 0% <= threshold 50% None

4. dynamiq/nodes/tools/file_tools.py:223

• Draft comment:
  Cleaned up inline comments in _process_file_with_converter; logic remains clear.
• Reason this comment was not posted:
  Confidence changes required: 0% <= threshold 50% None

5. dynamiq/nodes/tools/file_tools.py:318

• Draft comment:
  Extraneous comments around file processing setup in execute were removed, enhancing readability.
• Reason this comment was not posted:
  Confidence changes required: 0% <= threshold 50% None

Workflow ID: wflow_8xbIqIJMjMEefccL

^{You can customize by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

[ellipsis-dev]

Important

Looks good to me! 👍

Reviewed 237c704 in 1 minute and 0 seconds. Click for details.

• Reviewed 29 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 3 draft comments. View those below.
• Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. tests/unit/nodes/tools/test_file_tools.py:49

• Draft comment:
  Update assert to compare string directly (no decode) since FileReadTool now returns text.
• Reason this comment was not posted:
  Comment did not seem useful. Confidence is useful = 0% <= threshold 50% This comment is purely informative, explaining a change that was made. It doesn't provide a suggestion, ask for confirmation, or point out a potential issue. It seems to be explaining the rationale behind a change, which is not necessary according to the rules.

2. tests/unit/nodes/tools/test_file_tools.py:140

• Draft comment:
  Update integration test to check for string output (remove .decode()) for integration file content.
• Reason this comment was not posted:
  Comment looked like it was already resolved.

3. tests/unit/nodes/tools/test_file_tools.py:150

• Draft comment:
  Update assertion for reading file from a different storage: expect string output instead of decoding bytes.
• Reason this comment was not posted:
  Comment did not seem useful. Confidence is useful = 0% <= threshold 50% This comment seems to be purely informative, as it is just stating what the update is about without providing any actionable feedback or suggestions. It doesn't ask for confirmation or suggest any changes.

Workflow ID: wflow_tpC7qV9wIMbR3DPL

^{You can customize by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

[github-actions]

Coverage Report •

File	Stmts	Miss	Cover	Missing
dynamiq/nodes/converters
llm_text_extractor.py	192	114	40%	14, 66–68, 130, 140, 148–150, 162, 190–192, 194, 202, 226, 228–231, 238–239, 246–247, 249–252, 254–256, 258–259, 261–266, 268, 280, 282, 303, 305, 307–309, 311, 315, 339–340, 345, 347–348, 357, 361, 376, 382, 384–387, 401–402, 404–409, 422, 445–451, 454–455, 470–472, 524, 526–528, 556–558, 560, 568, 591, 593–596, 603–604, 611–612, 614–617, 619–620, 622–627, 629
dynamiq/nodes/tools
file_tools.py	215	69	67%	41, 44–45, 51–52, 59–60, 62–64, 174, 221–222, 245, 247, 250, 252–253, 263, 274–277, 280, 300, 326, 330, 333, 337, 342, 346–347, 353–355, 358, 360, 362, 364–366, 384, 386, 388–389, 391, 393–396, 398, 409, 423, 425, 427–428, 430, 432–435, 437, 448, 510–512, 562–564
dynamiq/utils
file_types.py	20	0	100%
TOTAL	18122	5817	67%

Tests	Skipped	Failures	Errors	Time
955	35 💤	0 ❌	0 🔥	3m 31s ⏱️

[ellipsis-dev]

Important

Looks good to me! 👍

Reviewed f7e4afc in 1 minute and 56 seconds. Click for details.

• Reviewed 62 lines of code in 2 files
• Skipped 0 files when reviewing.
• Skipped posting 2 draft comments. View those below.
• Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. dynamiq/nodes/tools/file_tools.py:68

• Draft comment:
  Ensure file type detection yields these new enum values. PPTX files may be detected as FileType.PRESENTATION rather than PPTX_PRESENTATION, so PPTXFileConverter might not be used.
• Reason this comment was not posted:
  Comment looked like it was already resolved.

2. dynamiq/utils/file_types.py:31

• Draft comment:
  Overlapping extensions for PPTX: 'pptx' appears in both FileType.PRESENTATION and FileType.PPTX_PRESENTATION, which can cause ambiguity during detection.
• Reason this comment was not posted:
  Comment was not on a location in the diff, so it can't be submitted as a review comment.

Workflow ID: wflow_WMyL9rUsiafLSDnc

^{You can customize by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

[vitalii-dynamiq]

@olbychos @tyaroshko @delamainer please review

[tyaroshko]

LGTM

(but we'll probably need to update this PR when we make dynamiq-extra public in order to avoid code duplication for file type mapping and detection)

[maksymbuleshnyi]

Example of trace

[olbychos]

Overall, lgtm; any other important file types worth adding (like excel,csv)?

[maksymbuleshnyi]

Overall, lgtm; any other important file types worth adding (like excel,csv)?

This is generally supported because we will fallback to pure content, which works well with csv

[ellipsis-dev]

Important

Looks good to me! 👍

Reviewed eaeafec in 37 seconds. Click for details.

• Reviewed 20 lines of code in 1 files
• Skipped 0 files when reviewing.
• Skipped posting 2 draft comments. View those below.
• Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

1. dynamiq/nodes/tools/file_tools.py:224

• Draft comment:
  Removed redundant comment; file.seek(0) clearly resets file pointer.
• Reason this comment was not posted:
  Confidence changes required: 0% <= threshold 50% None

2. dynamiq/nodes/tools/file_tools.py:312

• Draft comment:
  Removed redundant comment; the code is self-explanatory.
• Reason this comment was not posted:
  Confidence changes required: 0% <= threshold 50% None

Workflow ID: wflow_b0b3SsplkhockqmX

^{You can customize by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

[acoola]

LGTM. Please make sure that FileReadTool works fine with YAML load/dump, tracing and so on